GitLab

Sprout from master 1998-02-19 11:48:00 UTC Andrew Hodgkinson <ahodgkin@gitlab.riscosopen.org> 'Explained when to use the ahodgkin_133beta tag instead of ahodgkin_200release'
Cherrypick from master 1998-03-20 12:48:26 UTC Andrew Hodgkinson <ahodgkin@gitlab.riscosopen.org> '        Forgot this, too...':
    c/Multiuser

Precautions taken in table code to avoid divide by zero errors.

Framesets defining more frames than the frameset implies, for whatever
reason, should never cause a crash now - just report a 'Frames definition
is badly nested' error (if STRICT_PARSER defined). Along the same lines,
in a single document defining a set of nested frames, two /FRAMESETs in
succession could cause a fairly nasty crash. Fixed.

TIFFs do not get loaded by double-click now.

If holding down SHIFT to save a link contents to disc, you can also hold
down CTRL to bypass the cache (sets the browser's reloading flag). Noticed
when testing this that windows_create_browser didn't take account of the
Controls file 'UseSmall' entry when adjust-shift-clicking on links. It
does now.

There is now a complete and up to date set of interactive help messages
built into the Ursula, Customer and Phoenix build Res files.

WindowManager 3.98. Sets URI handler environment variables for
http, ftp and gopher.

Fixed 'doesn't reformat for unsized images' bug - only happens when the
deferred reformatter is enabled, hadn't remembered to check this in my
debug build where this is turned off. Oops.

Caret position in forms could get left behind despite attempts to rectify
this in v1.31 - now fixed (and faster, fortunately).

Hourglass + percentage displayed for History and Image History when
loading - it can take a while for very big histories, though in
practice you only see the visit history loading (has to do a lot
more work, and is thus quite a bit slower than the image history).

POST forms worked when targetted to frames, but were broken when
not targetted! (Forms data inherited from one browser to the same
browser; ended up freeing the flex block...). Fixed.

Now support 303 response code (redirect to GET). Treated as 301, i.e.
not support if STRICT_PARSER is ...

Machine hung during last check-in; trying again... Apologies for sending out such a huge log message *twice*.

Check for WindowManager 3.97 and ensure Unicode$Path is set in all
!Run[D] files, don't set the Alias$@PrintType_FF4 variable, and updated
Customer build ROM obey file variants. Various other changes to
the Run files for new module versions, updated paths to support
new positions of choices, hotlist and histories (see later), etc.

!Sprites[22] files hold small !app icons for some variants, and
an ic_browse sprite. Some variants now have a Sprites and Sprites22
file instead of just Sprites, with the former containing various
mode 12 or 15 specific sprites.

Text files dragged to the URL writable are treated as ANT URL files. This
relies on URLBarWrit (Toolbars.h) being a unique ID, which it should be,
but beware of the Hotlist and Choices numberspaces...

INPUT TYPE=BUTTON supported. Form items without a FORM tag are now shown
(as MSIE 4, but not NN 4).

'*', '-', '@', '_' and '.' are not escaped ...

Check for WindowManager 3.97 and ensure Unicode$Path is set in all !Run[D] files, don't set the Alias$@PrintType_FF4 variable, and updated

Customer build ROM obey file variants. Various other changes to
the Run files for new module versions, updated paths to support
new positions of choices, hotlist and histories (see later), etc.

!Sprites[22] files hold small !app icons for some variants, and
an ic_browse sprite. Some variants now have a Sprites and Sprites22
file instead of just Sprites, with the former containing various
mode 12 or 15 specific sprites.

Text files dragged to the URL writable are treated as ANT URL files. This
relies on URLBarWrit (Toolbars.h) being a unique ID, which it should be,
but beware of the Hotlist and Choices numberspaces...

INPUT TYPE=BUTTON supported. Form items without a FORM tag are now shown
(as MSIE 4, but not NN 4).

'*', '-', '@', '_' and '.' are not escaped when submitting forms now. The
Web interface to the IMDb now works.

INPUT TYPE=HIDDEN items will not affect t...

Browsers with use default colours set only get redrawn by choices when the default colours have actually been changed.

Not all resources are up to date in this check-in, and documentation both within source and stuff in 'Docs' is out of date or missing - I was very pushed for time on this one... Hopefully will do another 'tidy up' check-in before close on Friday; until then, beware of anything other than the Ursula (Desktop browser) build. Anyway...

Export As Draw done. As part of this, FONT SIZE and SUP / SUB adjustment
of font size is done in fm_token_font_info rather than fm_find_token_font.

Bullets and switches are plotted as indirected sprite items, rather than
indirected text + sprite items - this relied on being in a redraw loop to
pick the sprites up from the local pool (failed during printing).

URI files now have a LF line ending rather than CR... '*' recognised
as an empty field when loading via. RAM transfer now (only worked for
loading from disc before).

Gave placeholder bounding boxes 4 OS units more minimum extra gap
vertically and reduced horizontal addition to this value * 1.5, rather
than * 2 (see reformat_get_placeholder_size).

Background image tiling starts at ymax - h + 4, rather than ymax, so
there isn't the bottom line of pixels from the top tile always present
at the top of the window. The '+4' is for caution's sake.

Can now save a frame's HTML source, that of its parent or its ancestor,
and the same for the URI pointing to those documents - see Menus.h for
the relevant component IDs that should lead to the SaveFile dialogue.

Have hopefully fixed timeout = 0 values (i.e. 'forever') on things
like LinkTo; before, image fetches could override the state (so you'd
only ever see brief flickers of a given URL as the pointer went over
a link).

Save File dialogue will remember the state of option or radio buttons
for a given parent component origin and restore that state when the
dialogue is next opened from the same place (stops turning on saving
as a URL file also turning on 'save background images' for Draw file
export, etc.).

Use of a META tag to reload the *same* page now sets the 'reloading'
flag in the browser so that it doesn't go through a proxy - otherwise
pages which are meant to update periodically through client pull
don't work, as they keep coming out of the cache.

Can now handle images specifying just a width or height in the HTML
(other dimension is scaled accordingly, but note that the placeholder
size must still be 'dumb' until the image data comes in). An image
will now override an image history size entry for the same entry
with a different size.

Note that the 'Docs' documentation (possibly contrary to previous log messages) still hasn't been updated with the new SaveFile stuff, and function header comments have yet to be written for

SaveFile.c.

Fixed urlutils_pathname_to_url - OS_GSTrans doesn't like the same
input buffer as output buffer (fairly obvious, but I was living in
the hope of not having to duplicate the input path string).

Image history will not cache scrap file images anymore. Browser's
image handling library won't cross reference them either.

When dragging a single item from the hotlist to a Filer window, the
code now checks Ctrl - if held down, a URL file is written, else a
URI file. The state of any option buttons or radios in the last
Save File dialogue that was opened is no longer relevant...!

history_save_as_html could write out broken files when titles
for items in the History were not present. Fixed.

Hourglass switched on for hotlist and history saving as HTML;
hotlist_save_entries no longer closes the output file on error
(it should be done by the caller, since the caller is responsible
for opening the file and passing the FILE * pointer to
hotlist_save_entries).

Bug report prompted me to do this - now ignore <p> tags straight after <li> tags, so '<ul><li><p>Some text' works as the author (or automatic generator, more commonly) intended.

A <p> tag before a table will be acted upon now (it was ignored before).
Something like <li><p><table...> will give a line break and paragraph
space after the <li> despite the changes mentioned above, as in other browsers.
Both browsers differ from Navigator, where the table appears alongside
the <li> tag.

Multiple BR tags work as in MSIE / Navigator (they don't collapse to
zero height anymore).

For the above, line height (and so, <br><br> or <p> spacing) is now
calculated inside reformat_text_line_height, and is used with either
a token to get the text height at a given size (e.g. for general line
spacing or multiple BR tags) or NULL for normal size text (e.g. for
P spacing, though this value is in practice later scaled by 7/8ths in
reformat_check_height). Note this behaviour is exhibited by MSIE 4;
v3 gave two BRs the same spacing as one P (i.e. not font size
dependent).

Save dialogues can have an option button or couple of radios to
switch between various formats (e.g. URI or URL, sprite or original
format). See 'Docs.Notes' for more. The back-end to this is now
implemented (i.e. at the moment, save as URI or URL, save as original
image format). Note that when saving background images, a leafname
based on the original fetch URL is now offered rather than a generic
'Background'.

Internal URL scheme changed to be all lower case, so relativisation
through URL_Fetcher still makes sense... :-/

Holding down 'shift' when clicking on stop reverses the interpretation
of the Controls file 'StopWebServe' entry.

Fixed row / column count for exporting tables as text; fixed a few
bits and pieces of internal URL scheme stuff which got broken when
HTMLLib started using URL_Fetcher's relativisation; this includes
automatic expansion of system variables in pathnames, in
urlutils_pathname_to_url.

There are a few known significant problems with this code but it's being checked in so the Choices can be worked on. Note that Res files etc. are not up to date across all builds. Progress is as follows...

Bug report prompted me to do this - now ignore <p> tags straight after
<li> tags, so '<ul><li><p>Some text' works as the author (or automatic
generator, more commonly) intended.

A <p> tag before a table will be acted upon now (it was ignored before).
Something like <li><p><table...> will give a line break and paragraph
space after the <li> despite the changes mentioned above, as in other browsers.
Both browsers differ from Navigator, where the table appears alongside
the <li> tag.

Multiple BR tags work as in MSIE / Navigator (they don't collapse to
zero height anymore).

For the above, line height (and so, <br><br> or <p> spacing) is now
calculated inside reformat_text_line_height, and is used with either
a token to get the text height at a given size (e.g. for general line
spacing or multiple BR tags) or NULL for normal size text (e.g. for
P spacing, though this value is in practice later scaled by 7/8ths in
reformat_check_height). Note this behaviour is exhibited by MSIE 4;
v3 gave two BRs the same spacing as one P (i.e. not font size
dependent).

Save dialogues can have an option button or couple of radios to
switch between various formats (e.g. URI or URL, sprite or original
format). See 'Docs.Notes' for more. The back-end to this is now
implemented (i.e. at the moment, save as URI or URL, save as original
image format). Note that when saving background images, a leafname
based on the original fetch URL is now offered rather than a generic
'Background'.

Internal URL scheme changed to be all lower case, so relativisation
through URL_Fetcher still makes sense... :-/

Holding down 'shift' when clicking on stop reverses the interpretation
of the Controls file 'StopWebServe' entry.

Fixed row / column count for exporting tables as text; fixed a few
bits and pieces of internal URL scheme stuff which got broken when
HTMLLib started using URL_Fetcher's relativisation.

Made fm_get_string_width work for system font when the maximum width is less than the width of the first word.

Corrected width of wide characters to 16 dots not 17.

Fixed some glitches in the modified status bar code (e.g. status inference wasn't working properly).

Shift+Toggle size always opened to not obscure the icon bar. If CMOS is
set to always show the icon bar, then Shift+Toggle size is meant to
reverse sense and open to absolute full size. It now does this.

!Run[D] files now require URL 0.22, HTTP 0.64, FTP 0.22, Gopher 0.08
and File 0.32.

Minor overhaul of 'Docs.Notes' (with 'Docs.User' similarly updated) to
include a few new bits and pieces.

Timeout values of 0 now mean 'forever'. Plug-in status messages now have
their own timeout setting. New Messages token 'Actv' for extra info when
a status message times out and the status line drops to 'Viewing' or
'Ready' when there is still a Busy state registered by the Plug-in.

Small Fetch windows don't become visible briefly for Plug-in fetches
when the fetch is complete (for SeeFetches:no, where these windows
should be hidden all the time).

WebServe may be controlled through the Browse front-end via.
Wimp_MAppControl reason code 4 (Wimp_MAppControl_Configure). Note the
renaming internally of 'WebServ' or 'WebServe' has been changed to
'Proxy' (with lower case equivalents where relevant); proxy name
and starting details come from the Messages file 'ProxyName' and
'ProxyComm' tokens. More in 'Docs.Notes'.

Res files updated for image history and PRE/TT font aspect ratio
choices. 'Don't expire' labels changed to 'Expire' (oops, didn't
notice the change in the development Browse's Res file when I
updated all of the others until now...).

When printing, Wimp_PlotIcon appears to ignore R4 and R5, plots
'somewhere' on the page, and tries to get the sprite from the global
pool. This latter part normally causes an error to be raised; though
on my machine I just get a corrupted error message. The Desktop font
gets lost at this point too. All have yet to be fixed; as an
interim work around, Wimp_PlotIcon objects are never drawn at all
if the global 'printing' flag is set.

Message_PlugIn_Busy. Plug-in code more robust when given invalid
browser instance handles by the plug-in.

Line spacing is now calculated on the basis of the normal style base serif
font, with all other fonts being forced into that line height. This does
mean that an unusually tall (say) sans serif font may get clipped. It
appears to be the only way to get around wildly different baseline
depths returned from the font metrics - you can't work out line spacing
based on each different font style; the line spacing will vary.

Table heighting (as opposed to widthing...) improved considerably; rowspan
can no longer cause very tall cells in odd places. Having trouble getting
rid of the single pixel breaks between vertically adjacent cells, though
I've not tried too hard. Widthing, though, seems fairly badly broken at
present... :-/

Set/clear of page_is_text flag made more robust (it looked as though there
was the potential for this to get stuck in a set state, though I've never
see the front-end behave in a manner which indicates this is the case).
The reformatter will now decrease leading if this flag is set (plain text
pages look daft with a line spacing that is OK for 'rich' text pages).
No reformatting is done if the page width changes by dragging on the
resize icon, though toggle size / full screen will still reformat even
if the contents are only text (browser needs to sort out various width
flags at this point).

Cut down on excessive redrawing when reformatting due to a change in
window dimensions is not done. If display_width hasn't changed, then no
redraw is needed. If this causes redraw problems, then whatever is
changing display_width needs investigating. It shouldn't be kludged
(basically) by forcing a redraw instead of a reformat.

TT/PRE/etc. text can now have a non-100% aspect ratio. 80-90% looks
best (ArcWeb, for example, uses 86%). New option 'TTAspect' in
the Choices files.

Debug builds link to a non-debug Unicode library now; stops stderr
being dumped to the bottom left of the screen if you've not redirected
it in the Run file.

RISC OS 3.1 seems to need more initial WimpSlot than later OS versions.
The 64 deep nested table set gives a 'No stack for trap handler' error
(which it really means in this case!) without 800K, even though 3.71
is happy with just 640K. So, the !Run file checks if Boot$OSVersion
is exactly 300, 310 or 311, and sets the WimpSlot accordingly.

!Run[D] files now require latest fetcher module versions (URL 0.21,
File 0.31, HTTP 0.58).

i)  There can be two sets of gadgets, one set for page history and another set
    for image history.

ii) There can be one set of gadgets for both settings and two radio buttons to
    select between them.

Only one of these may be present at the same time.  Currently the radio button
version is the one that is in place.  To change to the other version delete the
radio buttons and copy the gadgets from the ChoicesTMP window into the
appropriate window.

Significant reduction in number of times reformatter is called during
a table widthing session results in a dramatic speed increase for nested
tables (e.g. 64 deep with a paragraph of text in the deepest table takes
about two seconds to load with near-instantaneous subsequent reformats,
as opposed to taking optimistically 28 days for a 32 deep table!).
This code relies on a very big stack of assumptions...

Save as plain text implemented.

Choices, Messages and Res files brought up to date where necessary.
Version number taken to 1.28, 04-Dec-97 (so I don't forget later,
basically...).

Forgot to add in ImgHistory.c / h last time. Note also that only the Browse and Ursula resources are up to date at this point - bits are missing from the other builds.

Image history - sizes of images are remembered for future reference (if the sizes aren't specified in the HTML, they can be looked for in the image history instead to minimise reformatting requirements). Choices file options ImageHistoryPath, ImageExpiryAge, ImageMaxSize and

SaveImageHistory added to support this (all have a direct analogy to
the global History items of the same name after removing 'Image').

Text area items are now the correct height for the number of rows
specified in the HTML, don't scroll a line too early, and have the
caret placed at the start of the text when moved into from above
or at the end when moved into from below (rather than always at
the end, which was behaviour inherited from single line writables,
where this is desirable). The top of the text area is aligned to
the top of any line of text on the same line, with the text area
dropping below the text, rather than the bottom of the area matching
the text baseline and extending upwards.

Filename extensions for download files are only stripped if an
attempt to convert to a RISC OS filetype through the Mime Mapper module
does not return an error or Data (0xffd) filetype. '.cgi' is a special
case which is always stripped.

Altered heading styles - all are bold, none are italic, H3 is as big
as H2 was, H2 is a little larger, and others scale down to H6 being
the size as normal text (but, as I say, bold).

Sorted out tables and forms. We can just look for tagno=TAG_FORM, and
get rid of the wobbly form_flag; also discarded the redundant
reprocess_table flag in fetch_preprocess_token. Two separate forms
in consecutive table cells (for example) which used to fail - they
were submitted as one single form - will now work correctly.

Fixed nasty bug in FontManage.c where font size 7 would intermittently
fail. The stupid font sizes that it could ask the Font Manager for may
be responsible for the occasional Font Manager crashes that have been
noticed. Array for font sizes was declared as [Limits_FontSizes], but
indices 1 to Limits_FontSizes are actually used - so needed a '+ 1'
in the declaration... :-/

Replaced the case insensitive string comparison functions in Utils.c
with more efficient versions by S.Brodie.

MAXLENGTH specifier in INPUT tags is now supported. If unspecified or
specifically zero, the browser overrides and allows any length.

All hotlist_find_match calls wrapped in '#ifndef REMOTE_HOTLIST' compile time option, for Customer builds.

Cross referencing of images when the background colour is different - should
not do it for backgrounds, for e.g. alpha transparent PNGs would then fail.
This was implemented, but it turns out it affects *all* cross referencing!
Fixed for just background images.

As described in Docs.TagNames, the tag name "ahodgkin_127beta2" associated
with this checkin applies to everything *except* Choices.c. v1.27 beta 2
was build with the Choices file tagged "ahodgkin_127beta". The confusion
arises because Choices.c got independently updated and checked back in
before I checked in the v1.27 beta 2 changes myself.

Choices dialogue only gets created the first time it is opened (ie. it is not deleted every time it is closed anymore).  Can be controlled by a #define

More tidying.  Made choices_seconds_to_typed_time a LOT smaller.  Made choices_return_appropriate_timetype do what it says it does (thankfully not a problem for the bete released version).

Adjust-click on forwards/backwards buttons did not work in the intermediate
build, and now does.

Adjust-click on form submission buttons where the submit method is POST
will now work correctly, as the new view inherits the extradata block
of the old. Failure of this is signalled through
make_no_fetch_memory_error, code 16 (see updated Errors file in Docs).
Note the name change of extradata in this context; more details below.

The passing of forms POST data and extra headers for sending in an
HTTP request via. the URL module was all done through pointers into
flex blocks, and was therefore very unsafe. All relevant functions now
use pointers to pointers to the data.

To remove confusion on the umpteen uses of the word 'extradata', the field
in browser_data has been renamed to 'post_data'. The 'extradata' parameters
used within FetchPage.c were all disconnected from the extra header info /
form POST data stuff that extradata is usually ...

This is an intermediate check-in to allow work on Choices for the new table options and History choices as detailed below. Res files are not up to date except where indicated and there are several known bugs that will be fixed before the 'final' v1.27 is created. Any work on resources should only be done for the testbed !Browse.

Client side image maps implemented. There is code to draw highlighted
borders in CSIM.c, but this is not wired in yet; other than that,
the implementation is functionally complete. As part of this, centralised
the fetching of a targetted URL taking into account user request of a
new view and full screen mode, in fetchpage_fetch_targetted. The forms
library now uses this too, so form buttons respond to both adjust-clicks
and TARGET attributes.

Fixed APPLET handling where '.class' isn't present in the CODE attribute.

Paragraphs squashed at the top of cells/pages - browser would insert white
space before.

Now append a ' ' to the end of History menu items to prevent the Wimp
thinki...

Stopped UNIFONT forcing system font use and added calls to Font_WideFormat.
Made sure the system font option is read before fm_init is called!

User agent string now set if Netscape faking is disabled (format used is
"Acorn Browse/1.26 (RISC OS 3.71)".

Removed 'The Onion' link from some Hotlist files. Time showed that the site wasn't entirely Family Friendly (TM)... Removed Shift Control (it's gone offline now), and corrected the Telegraph URL from ...the-telegraph...

to just ...telegraph...

StrongHelp manual generator now tries to work out the number of lines in
the table on the root page more intelligently.

Plug-in support extended to handle Shockwave. Quite a lot of underlying
work for this, including a pseudo cache system for temporary files.

Bug in redraw routines: they plotted background images within table
cells, when nothing should have been drawn. Oops. When background images
in cells are properly supported, this can get sorted out.

Tweaked vertical plot position of HRs (up to 4 OS units too high before due
to rounding).

Various Res file bits and pieces, Ursula build taken up to v1.26 beta
for release on the Browser site. *Note* not all resources are up to date!
This will be done after AW97.

Added support for background alpha PNGs.
Passed support_object flag through to HTMLLib.

Res files sorted out a bit, to add the Proxy option to the Choices dialogues
and fix the position of the menu popup in the Others pane.

Bogus comment in Global.h showed up when trying to build the Customer
variant (SINGLE_USER undefined); fixed.

Various fixes in the code required for Shockwave support (many invalid assumptions about the OBJECT tag structure). Not finished yet though.

Played around with signal handling 'n' all that to try and get sensible
errors reported for aborts etc. using code by KBracey. Need a ROM patch
for _swix on RISC OS 3.7 upwards (or a local fixed copy of _swix),
otherwise should sort out most of the problems seen (mostly 'no stack
for trap handler' errors and a distinct lack of a register dump).

Plug-In view on redraw / ASAP choices now work as they should (ASAP
was previously broken).

Version numbers in Messages incremented to 1.25 (23 Oct 1997).

For Plug-In support, created new source file RMA.c/h to handle RMA
claims and releases, in an attempt to ensure leakage never happens.
As such, this source is *slightly* more advanced than that in the 1.25
AW '97 build, though functionally identical. Well, that is, at least
you can confirm zero RMA leakage with '-debug LMem' in TRACE builds now!

Intend to do a consistency check on comment syntax in all sources
shortly - I need to rebuild my StrongHelp manuals; can't keep track
of all of these new functions...