[Edbrowse-dev] Debian

[Edbrowse-dev] Debian

[Karl Dahlke]

> I pushed the change for CURLOPT_VERIFYHOST.

Very good.
I hope it fixes some of the issues, which I don't seem to have over here,
so can't really test.

Karl Dahlke

Re: [Edbrowse-dev] Debian

[MENGUAL Jean-Philippe]

No. If your code builds, you should release. It will enable to upload in 
sid before freeze, then Adam or I will report a bug about the Adam's 
report, then we'll fix it in debian/rules. It seems to be a build 
problem due to debian package system. If we upload early and report 
quickly, maybe we can fix the bug before migration to jessie and speed 
up the bugfix.

Regards,

Le 25/09/2014 17:30, Chris Brannon a écrit :
> Adam Thompson <arthompson1990@gmail.com> writes:
>
>> However, it appears that the experimental debian package is unable to resolve
>> proxy server names, at least the one we're using where I work.
>> This works when I hand-compile (i.e.
>> use our makefile against the same libs as the debian package)
> Is this any reason to hold off on 3.5.1?
>
> -- Chris
> _______________________________________________
> Edbrowse-dev mailing list
> Edbrowse-dev@lists.the-brannons.com
> http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev
>


-- 

Jean-Philippe MENGUAL

accelibreinfo, votre partenaire en informatique adaptée aux déficients visuels

Mail: texou@accelibreinfo.eu

Site Web: http://www.accelibreinfo.eu

Re: [Edbrowse-dev] Debian

[Chris Brannon]

Adam Thompson <arthompson1990@gmail.com> writes:

> However, it appears that the experimental debian package is unable to resolve
> proxy server names, at least the one we're using where I work.
> This works when I hand-compile (i.e.
> use our makefile against the same libs as the debian package)

Is this any reason to hold off on 3.5.1?

-- Chris

Re: [Edbrowse-dev] Debian

[Adam Thompson]

[-- Attachment #1: Type: text/plain, Size: 447 bytes --]

On Tue, Sep 23, 2014 at 09:54:49PM -0700, Chris Brannon wrote:
> Karl Dahlke <eklhad@comcast.net> writes:
> 
> > Probably yes, as long as curl understands wild card certificates,
> 
> I'm fairly certain it does.  Anyway, I pushed the change for
> CURLOPT_VERIFYHOST.

Thanks.  It appears to work now, at least for my use-cases.
Also, thanks for getting this done so quickly,
it saved me a *lot* of time at work today.

Cheers,
Adam.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [Edbrowse-dev] Debian

[Chris Brannon]

Karl Dahlke <eklhad@comcast.net> writes:

> Probably yes, as long as curl understands wild card certificates,

I'm fairly certain it does.  Anyway, I pushed the change for
CURLOPT_VERIFYHOST.

-- Chris

Re: [Edbrowse-dev] Debian

[MENGUAL Jean-Philippe]

Hi,

Could you fill a bug on the Bug Tracker System, so that we could see 
what happens? We'd need to know if something is made in the Makefile 
(upstream) and not in the Deb package. Maybe it's due to a build 
dependency missing? What kind or dep could miss?

Oegards,

Le 23/09/2014 14:23, Adam Thompson a écrit :
> On Sun, Sep 21, 2014 at 11:20:43PM +0200, MENGUAL Jean-Philippe wrote:
>>>> So far no bugs has been reported about the experimental release of
>>>> Edbrowse. Do you plan a release soon? If yes, I'll package it in sid
>>>> so that it's in the next jessie. Otherwise, I could migrate the
>>>> current one from experimental to sid, but I don't like to import in
>>>> Debian dev releases of the packages.
> Unfortunately I was only able to install this on my work system yesterday,
> and so have only just found this bug.
> However, it appears that the experimental debian package is unable to resolve
> proxy server names, at least the one we're using where I work.
> This works when I hand-compile (i.e.
> use our makefile against the same libs as the debian package)
> so I think it's something to do with the way the program's being built for Debian.
> I can try and compose a proper bug report if that'll help,
> however I'm not sure of the best way to submit instructions to reproduce as I
> obviously can't talk too much about the nature of my workplace's proxy set up.
>
> Cheers,
> Adam.


-- 

Jean-Philippe MENGUAL

accelibreinfo, votre partenaire en informatique adaptée aux déficients visuels

Mail: texou@accelibreinfo.eu

Site Web: http://www.accelibreinfo.eu

Re: [Edbrowse-dev] Debian

[Adam Thompson]

[-- Attachment #1: Type: text/plain, Size: 966 bytes --]

On Tue, Sep 23, 2014 at 12:42:21PM -0400, Karl Dahlke wrote:
> > So the open question is whether CURLOPT_VERIFYHOST
> > should always match CURLOPT_VERIFYPEER?
> 
> Probably yes, as long as curl understands wild card certificates,
> wherein *.foobar.com matches www.foobar.com.
> I used such a certificate at my last job and it sure was convenient,
> and cheaper than buying a certificate for each subdomain.

I'd say the two should always match as well,
otherwise the behavior is counter-intuative (assuming curl treats the options separately)
particularly when using self-signed certs simply to get the encryption provided by ssl.
At the end of the day, when I switch off ssl verification I expect edbrowse to
behave like curl with the --insecure option, i.e.
do ssl regardless of whether the certificate can be verified.

Incidentally, the afore mentioned curl option works with the systems which
caused me to notice this bug.

Cheers,
Adam.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[Edbrowse-dev] Debian

[Karl Dahlke]

> So the open question is whether CURLOPT_VERIFYHOST
> should always match CURLOPT_VERIFYPEER?

Probably yes, as long as curl understands wild card certificates,
wherein *.foobar.com matches www.foobar.com.
I used such a certificate at my last job and it sure was convenient,
and cheaper than buying a certificate for each subdomain.

Karl Dahlke

Re: [Edbrowse-dev] Debian

[Chris Brannon]

Adam Thompson <arthompson1990@gmail.com> writes:

> Sorry for the late post (think I've missed the cut off),
> but I just found that I'm unable to disable ssl certificate verification with
> edbrowse when built on my work debian sid system.

Well, you missed the cutoff, but so did I.  If you think you can figure
out what is going on, I'm more than willing to hold off until we know
more.

I have a suspicion.  There's a second option related to SSL certificate
verification that is always set to true.  It is CURLOPT_VERIFYHOST.
CURLOPT_VERIFYPEER checks to see that the certificate is signed by a
recognized certificate authority.  CURLOPT_VERIFYHOST checks that the
hostname given in the certificate is the same as the hostname that you
are connecting to.  And these are separate checks.  For our purposes,
they probably don't need to be.  So the open question is whether
CURLOPT_VERIFYHOST should always match CURLOPT_VERIFYPEER?

-- Chris

as 

Re: [Edbrowse-dev] Debian

[Adam Thompson]

[-- Attachment #1: Type: text/plain, Size: 1064 bytes --]

On Sun, Sep 21, 2014 at 11:20:43PM +0200, MENGUAL Jean-Philippe wrote:
> >>So far no bugs has been reported about the experimental release of
> >>Edbrowse. Do you plan a release soon? If yes, I'll package it in sid
> >>so that it's in the next jessie. Otherwise, I could migrate the
> >>current one from experimental to sid, but I don't like to import in
> >>Debian dev releases of the packages.

Unfortunately I was only able to install this on my work system yesterday,
and so have only just found this bug.
However, it appears that the experimental debian package is unable to resolve
proxy server names, at least the one we're using where I work.
This works when I hand-compile (i.e.
use our makefile against the same libs as the debian package)
so I think it's something to do with the way the program's being built for Debian.
I can try and compose a proper bug report if that'll help,
however I'm not sure of the best way to submit instructions to reproduce as I
obviously can't talk too much about the nature of my workplace's proxy set up.

Cheers,
Adam.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [Edbrowse-dev] Debian

[Adam Thompson]

[-- Attachment #1: Type: text/plain, Size: 807 bytes --]

On Sun, Sep 21, 2014 at 10:53:19PM -0700, Chris Brannon wrote:
> Karl Dahlke <eklhad@comcast.net> writes:
> 
> > The only reason we didn't put a bow on 3.5.1 long ago is, waiting for debian.
> 
> Ok, sounds good to me.  If there are no objections by 07:00 GMT Tuesday,
> I'll tag 3.5.1 and upload the release.  Also I'll try to build some new
> static binaries in the next few days.

Sorry for the late post (think I've missed the cut off),
but I just found that I'm unable to disable ssl certificate verification with
edbrowse when built on my work debian sid system.
Neither the vs command or novs option in the config file work correctly.
Unfortunately, due to work commitments I can't do much debugging at work,
but something definitely isn't happy in our current code.

Cheers,
Adam.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [Edbrowse-dev] Debian

[Chris Brannon]

Karl Dahlke <eklhad@comcast.net> writes:

> The only reason we didn't put a bow on 3.5.1 long ago is, waiting for debian.

Ok, sounds good to me.  If there are no objections by 07:00 GMT Tuesday,
I'll tag 3.5.1 and upload the release.  Also I'll try to build some new
static binaries in the next few days.

-- Chris

Re: [Edbrowse-dev] Debian

[MENGUAL Jean-Philippe]

oops sorry, I used a wrong mail address to post. Thanks for forward.

Regards,

Le 21/09/2014 23:17, Chris Brannon a écrit :
> The following was posted to the list from an unsubscribed address by
> Jean-Phillippe Mengual.  I have the mailing list software set to
> autodiscard posts from non-subscribers, to prevent spam.  So I'll repost
> it here:
>
> edbrowse-dev-bounces@lists.the-brannons.com writes:
>
>> The attached message has been automatically discarded.
>>
>> Subject: Debian
>>
>> Hi,
>>
>> So far no bugs has been reported about the experimental release of
>> Edbrowse. Do you plan a release soon? If yes, I'll package it in sid
>> so that it's in the next jessie. Otherwise, I could migrate the
>> current one from experimental to sid, but I don't like to import in
>> Debian dev releases of the packages.
>>
>> What are plans?
>>
>> Regards,
> _______________________________________________
> Edbrowse-dev mailing list
> Edbrowse-dev@lists.the-brannons.com
> http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev
>


-- 

Jean-Philippe MENGUAL

accelibreinfo, votre partenaire en informatique adaptée aux déficients visuels

Mail: texou@accelibreinfo.eu

Site Web: http://www.accelibreinfo.eu

[Edbrowse-dev] Debian

[Karl Dahlke]

The only reason we didn't put a bow on 3.5.1 long ago is, waiting for debian.
Meantime we've put in a few more features and fixed a few bugs, which is fine,
but if they're happy then I'm happy.

Karl Dahlke

[Edbrowse-dev] Debian

[Chris Brannon]

The following was posted to the list from an unsubscribed address by
Jean-Phillippe Mengual.  I have the mailing list software set to
autodiscard posts from non-subscribers, to prevent spam.  So I'll repost
it here:

edbrowse-dev-bounces@lists.the-brannons.com writes:

> The attached message has been automatically discarded.
>
> Subject: Debian
>
> Hi,
>
> So far no bugs has been reported about the experimental release of
> Edbrowse. Do you plan a release soon? If yes, I'll package it in sid
> so that it's in the next jessie. Otherwise, I could migrate the
> current one from experimental to sid, but I don't like to import in
> Debian dev releases of the packages.
>
> What are plans?
>
> Regards,

Re: [Edbrowse-dev] debian

[Adam Thompson]

[-- Attachment #1: Type: text/plain, Size: 557 bytes --]

On Tue, Apr 01, 2014 at 06:24:27PM +0200, MENGUAL Jean-Philippe wrote:
> I'll give you a point this week-end, only time for me to be sure about the
> Debian-specific issue. I clarify the situation with devs about JS and
> libmozjs (Debian maintainers), then I do a feedback to you. I follow the
> issue :) But I need myself to understand all Debian requirements to give you
> the most useful info.

That's good to hear.  Any progress with this?

Cheers,
Adam.
PS: If you can give me info on how debian builds libmozjs I may be able to
assist with debugging

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [Edbrowse-dev] debian

[MENGUAL Jean-Philippe]

Hi,

I'll give you a point this week-end, only time for me to be sure about 
the Debian-specific issue. I clarify the situation with devs about JS 
and libmozjs (Debian maintainers), then I do a feedback to you. I follow 
the issue :) But I need myself to understand all Debian requirements to 
give you the most useful info.

Sincerely,


Le 01/04/2014 13:20, Adam Thompson a écrit :
> On Sat, Mar 08, 2014 at 09:45:09PM +0100, MENGUAL Jean-Philippe wrote:
>> Thanks for your news. I'll try upgrading the package in Sid and will do some
>> tests, but will tell you in detafls how things are in Debian. I give you
>> notices in a few days. Actually I've not followed the news here for some
>> time because I worked on other stuff, but just now I know there's an
>> evolution, I'll try and tell you all you need.
> Ok, any news on the edbrowse debian package?
>
> anything more we can do to help?
>
> Cheers,
> Adam.


-- 

Jean-Philippe MENGUAL

accelibreinfo, votre partenaire en informatique adaptée aux déficients visuels

Mail: texou@accelibreinfo.eu

Site Web: http://www.accelibreinfo.eu

Re: [Edbrowse-dev] debian

[Adam Thompson]

[-- Attachment #1: Type: text/plain, Size: 515 bytes --]

On Sat, Mar 08, 2014 at 09:45:09PM +0100, MENGUAL Jean-Philippe wrote:
> Thanks for your news. I'll try upgrading the package in Sid and will do some
> tests, but will tell you in detafls how things are in Debian. I give you
> notices in a few days. Actually I've not followed the news here for some
> time because I worked on other stuff, but just now I know there's an
> evolution, I'll try and tell you all you need.

Ok, any news on the edbrowse debian package?

anything more we can do to help?

Cheers,
Adam.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [Edbrowse-dev] debian

[MENGUAL Jean-Philippe]

Hi,

Thanks for your news. I'll try upgrading the package in Sid and will do 
some tests, but will tell you in detafls how things are in Debian. I 
give you notices in a few days. Actually I've not followed the news here 
for some time because I worked on other stuff, but just now I know 
there's an evolution, I'll try and tell you all you need.


Thanks
Le 08/03/2014 21:40, Karl Dahlke a écrit :
> I don't have a phone number for our contact over there, just email,
> as he originally emailed us to ask about edbrowse.
> So I sent him another email and explained the problem,
> and asked for all configuration settings, options, gcc version, etc.
> I hope he will get in contact with us.
>
> I added javaSessionFail() to the out of memory leg.
>
> Karl Dahlke
> _______________________________________________
> Edbrowse-dev mailing list
> Edbrowse-dev@lists.the-brannons.com
> http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev
>


-- 

Jean-Philippe MENGUAL

accelibreinfo, votre partenaire en informatique adaptée aux déficients visuels

Mail: texou@accelibreinfo.eu

Site Web: http://www.accelibreinfo.eu

[Edbrowse-dev] debian

[Karl Dahlke]

I don't have a phone number for our contact over there, just email,
as he originally emailed us to ask about edbrowse.
So I sent him another email and explained the problem,
and asked for all configuration settings, options, gcc version, etc.
I hope he will get in contact with us.

I added javaSessionFail() to the out of memory leg.

Karl Dahlke

[Edbrowse-dev] debian

[Karl Dahlke]

> Nope, web browsers aren't as forgiving as you thought.

Well the result of that is - we have to fill in all the missing variables,
all the objects thatmight be present in a full visual browser,
even if those objects and methods don't really do anything.
This of course is DOM, and would have to be done
before or after the big rewrite,
so the effort would not be in vein.
The more we know about DOM the better.

Or ... maybe ... there is a full open source js + dom we can leverage,
like Chrome, but it has to be completely disentangled from screens
and visual presentations.

Karl Dahlke

Re: [Edbrowse-dev] debian

[Chris Brannon]

Karl Dahlke <eklhad@comcast.net> writes:

> Run it under firefox and see what you get.
>
> alert("hello");
> var x = missingVariable;
> alert("world");

Well, I ran it under Chromium with ChromeVox, and Chromium refused to
proceed after the undefined variable error.  I also tried setting a
variable to the result of calling an undefined function:
var x = blahblahblah();
with the same result.  Nope, web browsers aren't as forgiving as you
thought.

-- Chris

[Edbrowse-dev] debian

[Karl Dahlke]

I spoke with Jean again - the next debian freeze is not for several months.
That's good; we are not "under the gun" to make these edbrowse changes.
We have a little time.
But I would like us to move forward still, making steady progress,
to make sure edbrowse is in the next debian release.
I don't think it was in the last one due to the js library incompatibility.
debian is the only commercial success I have enjoyed for any of my work,
ever, in the past 35 years.
And oddly enough I did not approach them at all;
they found edbrowse on their own, thought it a great program,
and wanted to package it.
But if I know they are putting it in their next release,
we can perhaps approach redhat and others and say,
"debian is including it, how about you?"
So we have some time, but we also want to move forward.
Not just to moz js 26, but also fixing some bugs that are in there.
Sadly, a few months is not enough time for the complete rewrite that I envision,
but time still to track down some of those seg faults that Adam referred
to, that could be due to js gc or some such.
I have to admit most seg faultts go away if you disable js.
Of course the bug that I fixed two days ago, an unterminated string
that could theoretically go on forever, might also seg fault,
though more often it just led to unpredictable results.
Anyways my hope is that 3.4.11 would be compatible
with the latest js, and farily stable.
Good enough for debian and perhaps other distros.

Something else I never figured out is how to get js
to march on even if it encounters an error.
It always stops executing, and that is often because of some visual
feature in the DOM that I did not implement.
The web page would likely work if it would just keep on going
after an error like this.
Or if I could resume execution after certain errors.
I could never find any way to do this, but I'll bet,
I haven't tested this yet but I'll bet that firefox,
using moz js, somehow continues to execute js even after an error.
It would be easy enough to test.
I make a web page containing this script.
Run it under firefox and see what you get.

alert("hello");
var x = missingVariable;
alert("world");

Short of the complete rewrite, the ability to march on after hitting an
undefined variable would make the biggest difference.

Meantime I will try to recall some of the sites I found
that produce a seg fault.

Thanks for all your work.

Karl Dahlke