maildev@lists.thunderbird.net

Thunderbird email developers

View all threads

Nightly crashing for the past two days

JK
Jonathan Kamens
Sun, May 26, 2019 10:42 PM

FYI, the last TB nightly build that doesn't crash on me is 2019-05-23. I
believe this is the culprit:
https://bugzilla.mozilla.org/show_bug.cgi?id=1554072. Oddly, though it's
marked P1, there has been no activity on it. I hope that's just because
of the long weekend and it will be addressed quickly after Memorial Day,
but shrug.

FYI, the last TB nightly build that doesn't crash on me is 2019-05-23. I believe this is the culprit: https://bugzilla.mozilla.org/show_bug.cgi?id=1554072. Oddly, though it's marked P1, there has been no activity on it. I hope that's just because of the long weekend and it will be addressed quickly after Memorial Day, but *shrug*.
A
ace
Sun, May 26, 2019 11:00 PM

Dňa 27. 5. 2019 o 0:42 Jonathan Kamens via Maildev napísal(a):

FYI, the last TB nightly build that doesn't crash on me is 2019-05-23. I
believe this is the culprit:
https://bugzilla.mozilla.org/show_bug.cgi?id=1554072. Oddly, though it's
marked P1, there has been no activity on it. I hope that's just because
of the long weekend and it will be addressed quickly after Memorial Day,
but shrug.

Hi, looks like a m-c bug. However m-c has made the app crash if
document.createElement is used from XUL files to create a XUL element
(not HTML). Either something in core/toolkit wasn't converted (and that
reported bug is real), or you may need to inspect your addons if they
may be calling that function wrongly.

See https://bugzilla.mozilla.org/show_bug.cgi?id=1553077 .

aceman

Dňa 27. 5. 2019 o 0:42 Jonathan Kamens via Maildev napísal(a): > FYI, the last TB nightly build that doesn't crash on me is 2019-05-23. I > believe this is the culprit: > https://bugzilla.mozilla.org/show_bug.cgi?id=1554072. Oddly, though it's > marked P1, there has been no activity on it. I hope that's just because > of the long weekend and it will be addressed quickly after Memorial Day, > but *shrug*. Hi, looks like a m-c bug. However m-c has made the app crash if document.createElement is used from XUL files to create a XUL element (not HTML). Either something in core/toolkit wasn't converted (and that reported bug is real), or you may need to inspect your addons if they may be calling that function wrongly. See https://bugzilla.mozilla.org/show_bug.cgi?id=1553077 . aceman
JK
Jörg Knobloch
Sun, May 26, 2019 11:51 PM

On 27/05/2019 01:00, ace wrote:

<pre class="moz-quote-pre" wrap="">Dňa 27. 5. 2019 o 0:42 Jonathan Kamens via Maildev napísal(a):
<pre class="moz-quote-pre" wrap="">FYI, the last TB nightly build that doesn't crash on me is 2019-05-23. I
believe this is the culprit:
<a class="moz-txt-link-freetext" href="https://bugzilla.mozilla.org/show_bug.cgi?id=1554072" moz-do-not-send="true">https://bugzilla.mozilla.org/show_bug.cgi?id=1554072</a>. Oddly, though it's
marked P1, there has been no activity on it. I hope that's just because
of the long weekend and it will be addressed quickly after Memorial Day,
but <b class="moz-txt-star"><span class="moz-txt-tag">*</span>shrug<span class="moz-txt-tag">*</span></b>.
<pre class="moz-quote-pre" wrap="">Hi, looks like a m-c bug. However m-c has made the app crash if
document.createElement is used from XUL files to create a XUL element
(not HTML). Either something in core/toolkit wasn't converted (and that
reported bug is real), or you may need to inspect your addons if they
may be calling that function wrongly.

See <a class="moz-txt-link-freetext" href="https://bugzilla.mozilla.org/show_bug.cgi?id=1553077" moz-do-not-send="true">https://bugzilla.mozilla.org/show_bug.cgi?id=1553077</a> .

Actually, it's bug https://bugzilla.mozilla.org/show_bug.cgi?id=1546338. Bob Clary also reported crashes that went away after updating an add-on. It doesn't crash for me after we (hopefully) fixed the last call sites.

So by themselves, Dailies after 2019-05-24 shouldn't crash.

Jörg.

JK
Jonathan Kamens
Mon, May 27, 2019 12:21 AM

Yup, that's what it was. Thanks to aceman and Jörg for the quick
explanation. I've submitted updates for my three add-ons which were
using document.createElement instead of document.createXULElement and
submitted issues / sent emails to the maintainers of several other
add-ons with the same problem.

I also submitted a PR to the add-on developer docs to alert other add-on
developers about this issue via that channel.

Incidentally, I noticed while looking at the add-ons impacted by this
that the next TB release is going to lose support for the second most
popular TB add-on, ImportExportTools, which has nearly 500,000 users,
unless someone either convinces the maintainer to upgrade it or takes
over maintenance of it. He has thus far proclaimed he won't be releasing
updates after TB60 (see https://freeshell.de/~kaosmos/index-en.html).

ImportExportTools provides extremely important functionality, as
evidenced by how many users it has. It will be a Big Deal if we lose it. :-/

  jik

On 5/26/19 7:51 PM, Jörg Knobloch wrote:

On 27/05/2019 01:00, ace wrote:

Dňa 27. 5. 2019 o 0:42 Jonathan Kamens via Maildev napísal(a):

FYI, the last TB nightly build that doesn't crash on me is 2019-05-23. I
believe this is the culprit:
https://bugzilla.mozilla.org/show_bug.cgi?id=1554072. Oddly, though it's
marked P1, there has been no activity on it. I hope that's just because
of the long weekend and it will be addressed quickly after Memorial Day,
butshrug.

Hi, looks like a m-c bug. However m-c has made the app crash if
document.createElement is used from XUL files to create a XUL element
(not HTML). Either something in core/toolkit wasn't converted (and that
reported bug is real), or you may need to inspect your addons if they
may be calling that function wrongly.

Seehttps://bugzilla.mozilla.org/show_bug.cgi?id=1553077  .

Actually, it's bug
https://bugzilla.mozilla.org/show_bug.cgi?id=1546338. Bob Clary also
reported crashes that went away after updating an add-on. It doesn't
crash for me after we (hopefully) fixed the last call sites.

So by themselves, Dailies after 2019-05-24 shouldn't crash.

Jörg.


Maildev mailing list
Maildev@lists.thunderbird.net
http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net

Yup, that's what it was. Thanks to aceman and Jörg for the quick explanation. I've submitted updates for my three add-ons which were using document.createElement instead of document.createXULElement and submitted issues / sent emails to the maintainers of several other add-ons with the same problem. I also submitted a PR to the add-on developer docs to alert other add-on developers about this issue via that channel. Incidentally, I noticed while looking at the add-ons impacted by this that the next TB release is going to lose support for the second most popular TB add-on, ImportExportTools, which has nearly 500,000 users, unless someone either convinces the maintainer to upgrade it or takes over maintenance of it. He has thus far proclaimed he won't be releasing updates after TB60 (see https://freeshell.de/~kaosmos/index-en.html). ImportExportTools provides extremely important functionality, as evidenced by how many users it has. It will be a Big Deal if we lose it. :-/   jik On 5/26/19 7:51 PM, Jörg Knobloch wrote: > On 27/05/2019 01:00, ace wrote: >> Dňa 27. 5. 2019 o 0:42 Jonathan Kamens via Maildev napísal(a): >>> FYI, the last TB nightly build that doesn't crash on me is 2019-05-23. I >>> believe this is the culprit: >>> https://bugzilla.mozilla.org/show_bug.cgi?id=1554072. Oddly, though it's >>> marked P1, there has been no activity on it. I hope that's just because >>> of the long weekend and it will be addressed quickly after Memorial Day, >>> but*shrug*. >> Hi, looks like a m-c bug. However m-c has made the app crash if >> document.createElement is used from XUL files to create a XUL element >> (not HTML). Either something in core/toolkit wasn't converted (and that >> reported bug is real), or you may need to inspect your addons if they >> may be calling that function wrongly. >> >> Seehttps://bugzilla.mozilla.org/show_bug.cgi?id=1553077 . > > Actually, it's bug > https://bugzilla.mozilla.org/show_bug.cgi?id=1546338. Bob Clary also > reported crashes that went away after updating an add-on. It doesn't > crash for me after we (hopefully) fixed the last call sites. > > So by themselves, Dailies after 2019-05-24 shouldn't crash. > > Jörg. > > > _______________________________________________ > Maildev mailing list > Maildev@lists.thunderbird.net > http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net
WM
Wayne Mery
Mon, May 27, 2019 2:08 AM

On 5/26/2019 8:21 PM, Jonathan Kamens via Maildev wrote:

He has thus far proclaimed he won't be releasing updates after TB60
(see https://freeshell.de/~kaosmos/index-en.html).

That's not actually what he says....

"Thunderbird 60 will be the last release which some of my extension
will be able to work with."

(Emphasis is mine. )

Also, last fall several add-on developers declared they would not be
updating their add-on for version but ultimately did so - when provided
additional information, some persuasion, time, etc.  I could be wrong
but I think he was one of them.

On 5/26/2019 8:21 PM, Jonathan Kamens via Maildev wrote: > He has thus far proclaimed he won't be releasing updates after TB60 > (see https://freeshell.de/~kaosmos/index-en.html). That's not actually what he says.... "Thunderbird 60 will be the last release which *_some_* of my extension will be able to work with." (Emphasis is mine. ) Also, last fall several add-on developers declared they would not be updating their add-on for version but ultimately did so - when provided additional information, some persuasion, time, etc.  I could be wrong but I think he was one of them.
WM
Wayne Mery
Mon, May 27, 2019 2:23 AM

On 5/26/2019 6:42 PM, Jonathan Kamens via Maildev wrote:

though it's marked P1, there has been no activity on it.

FYI P1 isn't blocker status nor is it "drop everything to fix this". 
Check the definitions at
https://mozilla.github.io/bug-handling/triage-bugzilla

On 5/26/2019 6:42 PM, Jonathan Kamens via Maildev wrote: > though it's marked P1, there has been no activity on it. FYI P1 isn't blocker status nor is it "drop everything to fix this".  Check the definitions at https://mozilla.github.io/bug-handling/triage-bugzilla
JK
Jörg Knobloch
Mon, May 27, 2019 5:55 AM

On 27/05/2019 02:21, Jonathan Kamens wrote:

I've submitted updates for my three add-ons which were using
document.createElement instead of document.createXULElement and
submitted issues / sent emails to the maintainers of several other
add-ons with the same problem.

Thanks, Jonathan.

The create{XUL}Element calls are not an issue for TB 68 and we don't
know what the future will hold for TB 76, most likely the complete
removal of XUL and XBL.

Jörg.

On 27/05/2019 02:21, Jonathan Kamens wrote: > I've submitted updates for my three add-ons which were using > document.createElement instead of document.createXULElement and > submitted issues / sent emails to the maintainers of several other > add-ons with the same problem. Thanks, Jonathan. The create{XUL}Element calls are not an issue for TB 68 and we don't know what the future will hold for TB 76, most likely the complete removal of XUL and XBL. Jörg.
JK
Jonathan Kamens
Mon, May 27, 2019 12:30 PM

I exchanged email with the author on April 10 and he said he was not
planning on attempting to port ImportExportTools to Thunderbird 68,
because of the effort that would be required.

I responded to him as shown below. He did not reply to my response:

Yes, I am well aware of the level of effort required to port
extensions to TB68, having just done the job for nine different
extensions and written a new extension for TB68 as well.

There was a big learning curve to get the first of my extensions
over the finish line, but after that it became for the most part
formulaic and got much easier.

If you'd like, I can take a look at ImportExportTools and give you
an idea of how complicated the port will be, or perhaps even help
with the necessary coding. Just point me at the source code if
you're interested.

Furthermore, given how extraordinarily popular your add-on is -- you
have almost 500,000 users, six times as many as my most popular
add-on! -- I suspect there may be Thunderbird developers willing to
pitch in and help with the porting for the sake of not losing the
functionality your add-on provides when TB68 ships. You could
subscribe to the "maildev" list
<http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net>
and inquire there.

I suspect TB68 is going to be the mainline release for at least a
year or two after it comes out, so I think it's worth a few days of
effort to port the legacy add-on even if it turns out it has to be
ported again to become a WebExtension.

I can't speak to why he seems to be unwilling to accept help. I suppose
I could just take the code and fix it and release a new version without
his consent, but given that he is reachable, as opposed to an add-on
maintainer having simply disappeared, it seems wrong somehow to do that
without his consent.

shrug I can't solve all the problems in the world, but I thought it
appropriate to mention this one to the folks here. Now I've done that.

  jik

On 5/26/19 10:08 PM, Wayne Mery wrote:

On 5/26/2019 8:21 PM, Jonathan Kamens via Maildev wrote:

He has thus far proclaimed he won't be releasing updates after TB60
(see https://freeshell.de/~kaosmos/index-en.html).

That's not actually what he says....

"Thunderbird 60 will be the last release which some of my
extension will be able to work with."

(Emphasis is mine. )

Also, last fall several add-on developers declared they would not be
updating their add-on for version but ultimately did so - when
provided additional information, some persuasion, time, etc.  I could
be wrong but I think he was one of them.

I exchanged email with the author on April 10 and he said he was not planning on attempting to port ImportExportTools to Thunderbird 68, because of the effort that would be required. I responded to him as shown below. He did not reply to my response: Yes, I am well aware of the level of effort required to port extensions to TB68, having just done the job for nine different extensions and written a new extension for TB68 as well. There was a big learning curve to get the first of my extensions over the finish line, but after that it became for the most part formulaic and got much easier. If you'd like, I can take a look at ImportExportTools and give you an idea of how complicated the port will be, or perhaps even help with the necessary coding. Just point me at the source code if you're interested. Furthermore, given how extraordinarily popular your add-on is -- you have almost 500,000 users, six times as many as my most popular add-on! -- I suspect there may be Thunderbird developers willing to pitch in and help with the porting for the sake of not losing the functionality your add-on provides when TB68 ships. You could subscribe to the "maildev" list <http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net> and inquire there. I suspect TB68 is going to be the mainline release for at least a year or two after it comes out, so I think it's worth a few days of effort to port the legacy add-on even if it turns out it has to be ported again to become a WebExtension. I can't speak to why he seems to be unwilling to accept help. I suppose I could just take the code and fix it and release a new version without his consent, but given that he is reachable, as opposed to an add-on maintainer having simply disappeared, it seems wrong somehow to do that without his consent. *shrug* I can't solve all the problems in the world, but I thought it appropriate to mention this one to the folks here. Now I've done that.   jik On 5/26/19 10:08 PM, Wayne Mery wrote: > On 5/26/2019 8:21 PM, Jonathan Kamens via Maildev wrote: >> He has thus far proclaimed he won't be releasing updates after TB60 >> (see https://freeshell.de/~kaosmos/index-en.html). > > That's not actually what he says.... > > "Thunderbird 60 will be the last release which *_some_* of my > extension will be able to work with." > > (Emphasis is mine. ) > > Also, last fall several add-on developers declared they would not be > updating their add-on for version but ultimately did so - when > provided additional information, some persuasion, time, etc.  I could > be wrong but I think he was one of them. >
WM
Wayne Mery
Mon, May 27, 2019 1:31 PM

Thanks for reaching out to him and offering to help.  Actions like that
have been very helpful to other authors, and generally what we need to
make add-ons more of the community it should be.

IIRC He is known for not responding, so it would not be surprising to
never get a response.

On 5/27/2019 8:30 AM, Jonathan Kamens wrote:

I exchanged email with the author on April 10 and he said he was not
planning on attempting to port ImportExportTools to Thunderbird 68,
because of the effort that would be required.

I responded to him as shown below. He did not reply to my response:

 Yes, I am well aware of the level of effort required to port
 extensions to TB68, having just done the job for nine different
 extensions and written a new extension for TB68 as well.

 There was a big learning curve to get the first of my extensions
 over the finish line, but after that it became for the most part
 formulaic and got much easier.

 If you'd like, I can take a look at ImportExportTools and give you
 an idea of how complicated the port will be, or perhaps even help
 with the necessary coding. Just point me at the source code if
 you're interested.

 Furthermore, given how extraordinarily popular your add-on is --
 you have almost 500,000 users, six times as many as my most
 popular add-on! -- I suspect there may be Thunderbird developers
 willing to pitch in and help with the porting for the sake of not
 losing the functionality your add-on provides when TB68 ships. You
 could subscribe to the "maildev" list
 <http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net>
 and inquire there.

 I suspect TB68 is going to be the mainline release for at least a
 year or two after it comes out, so I think it's worth a few days
 of effort to port the legacy add-on even if it turns out it has to
 be ported again to become a WebExtension.

I can't speak to why he seems to be unwilling to accept help. I
suppose I could just take the code and fix it and release a new
version without his consent, but given that he is reachable, as
opposed to an add-on maintainer having simply disappeared, it seems
wrong somehow to do that without his consent.

shrug I can't solve all the problems in the world, but I thought it
appropriate to mention this one to the folks here. Now I've done that.

  jik

On 5/26/19 10:08 PM, Wayne Mery wrote:

On 5/26/2019 8:21 PM, Jonathan Kamens via Maildev wrote:

He has thus far proclaimed he won't be releasing updates after TB60
(see https://freeshell.de/~kaosmos/index-en.html).

That's not actually what he says....

"Thunderbird 60 will be the last release which some of my
extension will be able to work with."

(Emphasis is mine. )

Also, last fall several add-on developers declared they would not be
updating their add-on for version but ultimately did so - when
provided additional information, some persuasion, time, etc.  I could
be wrong but I think he was one of them.

Thanks for reaching out to him and offering to help.  Actions like that have been very helpful to other authors, and generally what we need to make add-ons more of the community it should be. IIRC He is known for not responding, so it would not be surprising to never get a response. On 5/27/2019 8:30 AM, Jonathan Kamens wrote: > > I exchanged email with the author on April 10 and he said he was not > planning on attempting to port ImportExportTools to Thunderbird 68, > because of the effort that would be required. > > I responded to him as shown below. He did not reply to my response: > > Yes, I am well aware of the level of effort required to port > extensions to TB68, having just done the job for nine different > extensions and written a new extension for TB68 as well. > > There was a big learning curve to get the first of my extensions > over the finish line, but after that it became for the most part > formulaic and got much easier. > > If you'd like, I can take a look at ImportExportTools and give you > an idea of how complicated the port will be, or perhaps even help > with the necessary coding. Just point me at the source code if > you're interested. > > Furthermore, given how extraordinarily popular your add-on is -- > you have almost 500,000 users, six times as many as my most > popular add-on! -- I suspect there may be Thunderbird developers > willing to pitch in and help with the porting for the sake of not > losing the functionality your add-on provides when TB68 ships. You > could subscribe to the "maildev" list > <http://lists.thunderbird.net/mailman/listinfo/maildev_lists.thunderbird.net> > and inquire there. > > I suspect TB68 is going to be the mainline release for at least a > year or two after it comes out, so I think it's worth a few days > of effort to port the legacy add-on even if it turns out it has to > be ported again to become a WebExtension. > > I can't speak to why he seems to be unwilling to accept help. I > suppose I could just take the code and fix it and release a new > version without his consent, but given that he is reachable, as > opposed to an add-on maintainer having simply disappeared, it seems > wrong somehow to do that without his consent. > > *shrug* I can't solve all the problems in the world, but I thought it > appropriate to mention this one to the folks here. Now I've done that. > >   jik > > On 5/26/19 10:08 PM, Wayne Mery wrote: >> On 5/26/2019 8:21 PM, Jonathan Kamens via Maildev wrote: >>> He has thus far proclaimed he won't be releasing updates after TB60 >>> (see https://freeshell.de/~kaosmos/index-en.html). >> >> That's not actually what he says.... >> >> "Thunderbird 60 will be the last release which *_some_* of my >> extension will be able to work with." >> >> (Emphasis is mine. ) >> >> Also, last fall several add-on developers declared they would not be >> updating their add-on for version but ultimately did so - when >> provided additional information, some persuasion, time, etc.  I could >> be wrong but I think he was one of them. >>