This legacy header appears to be unused. Removing its includes
avoids numerous warnings when compiling with musl libc:
/usr/include/sys/signal.h:1:2: warning: #warning redirecting incorrect #include <sys/signal.h> to <signal.h> [-Wcpp]
Option document.browse.preferred_document_width controls the
width of the document, so that documents are rendered with narrower
width than screen width. Makes it easier to read paragraphs.
Patch originally from Shalon Wood <dstar@pele.cx>, see bug #1063.
Instead of using max_document_width as the hard limit to the document
width, it uses a soft limit, where if the document does not fit (due to
tables, etc.), then larger width is used. This reduces the need for
horizontal scrolling for wide documents.
Also added toggle-document-width action to toggle between preferred
width and full screen width. This is bound to 'M' by default. Initial
toggle status is determined by document.browse.use_preferred_document_width
option.
During dumps, document.dump.width option is still used. Perhaps we
should consolidate document.dump.width option with
document.browse.preferred_document_width ?
The color is controlled by
document.browse.links.active_link.insert_mode_colors.background
document.browse.links.active_link.insert_mode_colors.text
Also avoid overloading local variable "i" in get_current_link().
These settings are specified by
document.colors.link_number
document.colors.use_link_number_color
The latter setting determines whether the color is used when document
colors are being used.
See bug #1050.
If this function is available and gnutls is used,
and connection.ssl.trusted_ca_file is "",
gnutls_certificate_set_x509_system_trust is called
to load certificates.
When a mailcap entry with a testcommand containing %s is encountered, it is ignored. The
entry was freed, but not set to NULL, leading to a segfault later on.
This prevents links with null selectors being interpreted as pointing to
a (likely non-existent) file gopher://server/1
An example of a page with such links: gopher://gopher.floodgap.com/1/new
If there are isolated combining characters, e.g. at the beginning of a paragraph
or table cell:
– if it’s not the first screen column, combine them with whatever character is
printed to their left;
– otherwise, add a no-break space as the base character.
Previously, such combining characters were combined with the last letter
displayed, i.e. the last letter of the previous paragraph or cell.
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
without waiting for the next non zero width character. This way combining
characters at the end of the document are displayed.
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
Otherwise, there are unnecessary spaces at the end of lines in tables containing
combining characters.
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
A combining character sequence where the base character is a space remained
recorded as a space although the initial space was replaced with an internal
code corresponding to the combined character. This caused an internal error when
ELinks tried to split the line at that place and did not find the space.
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
Otherwise, the verification fails when using a HTTPS proxy because ELinks checks
whether the proxy matches the certificate.
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
I didn’t respect the right order when I added the functions to move up and down
by half a page a few years ago.
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
Use the same functions as for the list-style property since only the "type" part
of the list-style property is supported at this stage.
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
* Rename P_STAR as P_DISC and P_PLUS as P_SQUARE.
* Delete P_NONE because it was used only as the default flag in init_html_parser
and a list with P_NONE then got bullets, so instead use P_DISC by default (as
per the CSS specification), and P_NO_BULLET for lists with no bullets.
* Use as bullets the characters:
- U+25E6 WHITE BULLET for the circle style;
- U+25AA BLACK SMALL SQUARE (alias square bullet) for the square style;
- U+2022 BULLET for the disc style (default).
Signed-off-by: Fabienne Ducroquet <fabiduc@gmail.com>
By default some Python warning messages would be written to standard error
by the interpreter. To prevent these warnings from making a mess of the
ELinks screen, all warnings were turned into exceptions so they could be
caught and displayed through the usual report_scripting_error() mechanism.
With Python 3.2, this approach backfires: A new class of ResourceWarnings
that are filtered by default (and hence would *not* have been written to
standard error) are now turned into exceptions, and these exceptions can't
be caught because they're emitted from the interpreter's cleanup code. As
a result, the uncaught exceptions would make a mess of the ELinks screen.
The new solution is to replace Python's standard library function
warnings.showwarning() with one that turns warning messages into exceptions.
This means we'll wait until a warning would have been written to standard
error before turning it into an exception, so other warnings that would
never have reached that point because they're filtered will remain unseen.
(The behavior of warning messages is described in the documentation for
the "warnings" module from Python's standard library.)
There was hardcoded 4096 for threads size, but below there was:
assertm(fd >= 0 && fd < FD_SETSIZE,
"get_handler: handle %d >= FD_SETSIZE %d",
fd, FD_SETSIZE);
if_assert_failed return NULL;
which fails for fd > 1024 (1024 was previous value of FD_SETSIZE)
DELETED:
* dumb {b}, smart {bb, bb_fr_en, bb_en_fr}: Redirects to Microsoft
Translator, which I can't get to work without JavaScript. Deleted.
* dumb {pyhelp}, smart {py, pydev}: Deleted as recommended by the
author of the CGI script.
* dumb {pyvault, lyrics}, smart {pyvault}: Can't find a new URL for the
service. Deleted.
* smart {gd}: Google Directory has been shut down. Deleted.
* smart {sd, sdc, sdu, sdp, sdj, whatis}: These don't seem to work
without JavaScript. Deleted.
* smart {id, draft}: The search still exists but I can't get it to
take the words from the URL. Deleted.
UPDATED:
* dumb {cia}: Had changed its URL, and the service is down, but they
intend to restore it. Updated and kept.
* dumb {lua}: ELinks no longer supports Lua 4.0. Changed to 5.1 as
installed by Debian.
* smart {cliki, foldoc, gwho, gwhat, gwhere, gwhen, a, imdb, wn, fsd,
rfcs, cr}: Updated URLs.
Each C function that can be called as a method of an ECMAScript object
is typically listed in a spidermonkeyFunctionSpec array and has a
comment that shows the name of the array and the name of the method.
For example, elinks_alert has a comment /* @elinks_funcs{"alert"} */
because elinks_funcs[] contains an element { "alert", elinks_alert, 1 }.
When some of those functions were split into outer and inner functions
for SpiderMonkey 1.8.5 compatibility, the comments were attached to
the inner functions, which contain the bulk of the code. Move the
comments to the outer functions, to which the array elements point.
Partially revert commits
a1c5fe51 (2011-05-08) bookmark_folder_get_property: xulrunner-2.0 fix
e86ec567 (2012-03-04) Compilation fixes (--enable-debug)
jsval_to_bookmark_string once again takes jsval val, rather than jsid id.
This way, bookmark_set_property does not have to call JS_ValueToId,
which would needlessly intern the new string value of the property,
i.e. the title or the URI. Instead, bookmark_folder_get_property
has to call JS_IdToValue.
forms_item declared a variable (jsval val), passed its address to
forms_item2, and set it as the return value. However, forms_item2
could return without initializing the jsval, especially if given too
many arguments. Fix by initializing to JSVAL_VOID right away.
I'm not sure that is the correct value to return in such cases,
but at least it's better than risking a crash.
Likewise in form_elements_item and form_elements_namedItem.
Fix this warning:
Generating docs for page todo...
todo:12: warning: explicit link request to 'url' could not be resolved
The #url syntax was supposed to refer to the url member of the
containing structure, but it apparently doesn't work when @todo
copies the paragraph to a separate page.
Doxygen warned that <uri> is an unsupported xml/html tag.
It was not intended as a tag though; rather as a placeholder
akin to <var>uri</var>. Escape with a backslash.
In almost all calls to parse_header, the second argument is a string
literal: parse_header(head, "Charset", NULL) for example.
Of course, parse_header does not write to that string. Accordingly,
make the parameter point to const, so calls that use a variable rather
than a string literal can make that variable point to const too.
Leaving the other parameters non-const for now.
http_negotiate_parse_data passed &token->length as the int *outlen
parameter of base64_decode_bin, which stores an int at that location.
However, gss_buffer_desc::length is size_t in all implementations that
I checked: MIT Kerberos Version 5 Release 1.10, libgssglue 0.4, and
GNU GSS 1.0.2. This mismatch could cause the build to fail:
.../src/protocol/http/http_negotiate.c: In function ‘http_negotiate_parse_data’:
.../src/protocol/http/http_negotiate.c:173:2: error: passing argument 3 of ‘base64_decode_bin’ from incompatible pointer type [-Werror]
In file included from .../src/protocol/http/http_negotiate.c:30:0:
.../src/util/base64.h:8:16: note: expected ‘int *’ but argument is of type ‘size_t *’
On 64-bit big-endian hosts, it might also cause the GSSAPI
implementation to read too much data from memory and disclose it to
some network server, or crash ELinks.
(cherry picked from elinks-0.12 commit d33c807dd9)
unregister_options() requires as a sentinel an instance of struct
option where option.name is NULL. However, the NULL_OPTION_INFO macro
used for these sentinels actually initializes a struct option_init
instead. Make register_options() overwrite the NULL_OPTION_INFO with
a sentinel in the correct format. This probably makes a difference
only on platforms where null pointers don't have all bits zero.
(cherry picked from elinks-0.12 commit 8ac10e00d4)
l_set_option() was passing the address of an int to
option_types[OPT_INT].set and option_types[OPT_BOOL].set.
That looks correct but is not: both function pointers
point to num_set(), which actually reads *(long *) str.
Change l_set_option() to pass the address of a long instead,
and add comments about this dependency.
(cherry picked from elinks-0.11 commit 8766e3829f)
(cherry picked from elinks-0.12 commit 0797f04921)
