Bugs In Previous Grail Releases
Bugs in Grail 0.5
There were bugs in Grail 0.5, but we didn't bother to keep track
Known Bugs in Grail 0.4
- When using (recent versions of) PIL, a traceback ensues whenever
Grail tries to load an image.
Known Bugs in Grail 0.3
Here is a list of known bugs in Grail 0.3. We plan to fix all of
them in future versions. If you're using Grail 0.3 to read this page,
you will be able see to which bugs have been fixed on the development
version (meaning with the final 0.3 release): the items will be
bulleted with an open square instead of round bullet.
See also the list of fixed bugs that were
present in Grail 0.2.
These bugs have all been fixed in the final release of Grail 0.3.
- We've received many reports of Grail hanging on
64-bit hardware, in particular DEC Alpha using either Linux or DEC
Unix. This appears to be a bug in Tk. We received a description of
the problem and a fix for Tk from
- As distributed, Grail does not work correctly with
some HTTP/1.1 servers. Here's a patch.
- The README file is wrong about the requirement for
Python 1.3 -- you must use at least 1.4.
- Crash reading Netscape version 1 history files.
Here's a patch.
- Crash in table code on
- Attributes like
HREF=~username are not
parsed the same as other browsers.
- Unvisited anchors are purple instead of blue.
- Content elements of list items (LI, LH, DD, DT)
were not always closed correctly.
- The active link isn't re-colored on the button
- Selecting the label format for <OL>,
<UL>, and <LI> was not checking the type of the list.
- Printing images that require a page break to be
inserted messes up the page length.
- Partially loaded images are not displayed (this would require changes
- No built-in support for these protocols: gopher, wais, telnet.
- No in-line scaled images. (Fix: This is supported
when the Python Imaging Library is available and enabled, and in
- Sometimes images don't get loaded with no indication as to the
- Memory leaks.
- It's too slow (faster than Grail 0.2, but still too slow...).
- See also a note on the
current tables implementation.
- Printing of images does not detect whether
installed; this should be detected and reported to the user in the
printing dialog if the Python Imaging Library is not used.
- Table markup discarded when printing; this results in seriously
mangled tables most of the time.
- Printing doesn't support any extensibility mechanisms
other than adding and changing dingbat images.
<BODY BACKGROUND="image-uri"> is not
supported. This isn't likely to change until Tk supports background
- Applets can't use modules that are imported as shared libraries.
- Applet code is executed synchronosly, so any blocking I/O they
might do (such as loading URLs) blocks the entire browser.
- Applets can read arbitrary files (they can only write files in
$TMPDIR or /usr/tmp though).
- Applets can open arbitrary sockets.
- Applets still have access to some browser internals and can modify
- Applets can quit the browser (in a number of ways!).
- There is no limit to the resources used by an Applet.
- There could be an applet debugger.
- The save dialog for files of unknown type should
display the URL and file type, and suggest a filename. (Fix: the MIME
type is displayed at the top of the dialog.)
- <FIG> support.
- Support for the Secure HTTP protocol (https: URLs).
(This bug was reported for Grail 0.2 but could occur with Grail 0.3 too.)
- The documentation is still incomplete.
- It is no longer possible to provide "binary" versions of Grail,
due to incompatibilities between the Python freeze utility and the
package implementation ("ni.py").
(These bugs were reported for Grail 0.2 but could occur with Grail 0.3
- Sometimes, Grail seems to "just hang" on Linux, especially when
built from the source. A solution seems to be ro run the regen.py
script in the Lib/linux1 directory; there seem to be different
constants in the header files depending on which Linux version you are
There seem to be some problems with the GIF reading code in Tk.
- We have received some reports of Solaris 2.4
on Sparc and Intel/x86 hanging (with spinning busy icon) forever
during load of a page. Sometimes using Reload fixes this,
sometimes not. We have reason to believe that there's a patch to
Solaris 2.4 (for both Sparc and Intel) that fixes this. The closest
description of the patch is BugId #1186224, having something to do
with "select hanging in NON-BLOCKING mode."
This bug appears to be fixed with these jumbo kernel patches: PatchId
#101945 for Sparc and #101946 for Intel. For Sparc you need at least
level 19, and for Intel, you need level 29. As of 16-Nov-1995,
101945-34 is the latest for Sparc, and 101946-29 is the latest for
x86. With these patches installed, we have not experienced the
We have not experienced this problem with Solaris 2.5 on either Sparc
We now supply source patches for Tk for both
- Some GIF images are displayed incorrectly (e.g. some in the
October 95 issue of D-Lib magazine).
- Certain GIF images cause Python to dump core in the GIF or PPM (!)