Hi - I've got both Zinc and Zodiac loaded into a Pharo 2 image. When I do a ZnClient #get: to an http connection I'm getting the dnu mentioned in the subject of this email. The message send that causes it is in ZnStringEntity>>#readFrom: ZnNetworkingUtils default socketStreamClass prints: 'ZdcSocketStream' Changing that to 'SocketStream' like so: ZnNetworkingUtils default socketStreamClass: SocketStream makes everything work . Just wanted to let you know Paul
Hi Paul, On 30 Jun 2013, at 19:38, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi -
I've got both Zinc and Zodiac loaded into a Pharo 2 image. When I do a ZnClient #get: to an http connection I'm getting the dnu mentioned in the subject of this email.
The message send that causes it is in ZnStringEntity>>#readFrom:
ZnNetworkingUtils default socketStreamClass prints: 'ZdcSocketStream'
Changing that to 'SocketStream' like so:
ZnNetworkingUtils default socketStreamClass: SocketStream
makes everything work .
Just wanted to let you know
Paul
Thanks a lot for reporting this issue. It is a bit weird though: I normally don't use #position on principle, but apparently I did in some recent optimisation code, I also vaguely remember adding #position somewhere.. Furthermore, all unit tests are green on both 2.0 and latest 3.0 with ZdcSocketStream as default. Step one would be to add a unit test. Is it possible for you to give me the URL or a similar one exposing the problem - it would have to be related to non-ASCII, non-Latin-1 characters I guess. Thanks, Sven -- Sven Van Caekenberghe Proudly supporting Pharo http://pharo.org http://association.pharo.org http://consortium.pharo.org
Hi Again, Paul, It was a typo ;-) I added two new unit test as well. Here are the relevant commits: === Name: Zinc-HTTP-SvenVanCaekenberghe.372 Author: SvenVanCaekenberghe Time: 30 June 2013, 9:06:10.101049 pm UUID: 98ea090f-27d4-498e-8111-95e954d010a9 Ancestors: Zinc-HTTP-SvenVanCaekenberghe.371 Fixed a typo in ZnStringEntity>>#readFrom: (thanks Paul DeBruicker for reporting the issue) === Name: Zinc-Tests-SvenVanCaekenberghe.191 Author: SvenVanCaekenberghe Time: 30 June 2013, 9:07:42.519249 pm UUID: 25345906-5574-4fd3-a5d9-eeee711fdf58 Ancestors: Zinc-Tests-SvenVanCaekenberghe.190 Added ZnClientTests>>testGet[Large]WideStringHTML === You'll have to load #bleedingEdge to get these. Thanks again for testing the latest version and reporting the issue. Sven On 30 Jun 2013, at 20:19, Sven Van Caekenberghe <sven@stfx.eu> wrote:
Hi Paul,
On 30 Jun 2013, at 19:38, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi -
I've got both Zinc and Zodiac loaded into a Pharo 2 image. When I do a ZnClient #get: to an http connection I'm getting the dnu mentioned in the subject of this email.
The message send that causes it is in ZnStringEntity>>#readFrom:
ZnNetworkingUtils default socketStreamClass prints: 'ZdcSocketStream'
Changing that to 'SocketStream' like so:
ZnNetworkingUtils default socketStreamClass: SocketStream
makes everything work .
Just wanted to let you know
Paul
Thanks a lot for reporting this issue.
It is a bit weird though: I normally don't use #position on principle, but apparently I did in some recent optimisation code, I also vaguely remember adding #position somewhere..
Furthermore, all unit tests are green on both 2.0 and latest 3.0 with ZdcSocketStream as default.
Step one would be to add a unit test.
Is it possible for you to give me the URL or a similar one exposing the problem - it would have to be related to non-ASCII, non-Latin-1 characters I guess.
Thanks,
Sven
-- Sven Van Caekenberghe Proudly supporting Pharo http://pharo.org http://association.pharo.org http://consortium.pharo.org
https://pharo.fogbugz.com/default.asp?11088 On Jun 30, 2013, at 9:11 PM, Sven Van Caekenberghe <sven@stfx.eu> wrote:
Hi Again, Paul,
It was a typo ;-) I added two new unit test as well.
Here are the relevant commits:
=== Name: Zinc-HTTP-SvenVanCaekenberghe.372 Author: SvenVanCaekenberghe Time: 30 June 2013, 9:06:10.101049 pm UUID: 98ea090f-27d4-498e-8111-95e954d010a9 Ancestors: Zinc-HTTP-SvenVanCaekenberghe.371
Fixed a typo in ZnStringEntity>>#readFrom: (thanks Paul DeBruicker for reporting the issue) === Name: Zinc-Tests-SvenVanCaekenberghe.191 Author: SvenVanCaekenberghe Time: 30 June 2013, 9:07:42.519249 pm UUID: 25345906-5574-4fd3-a5d9-eeee711fdf58 Ancestors: Zinc-Tests-SvenVanCaekenberghe.190
Added ZnClientTests>>testGet[Large]WideStringHTML ===
You'll have to load #bleedingEdge to get these.
Thanks again for testing the latest version and reporting the issue.
Sven
On 30 Jun 2013, at 20:19, Sven Van Caekenberghe <sven@stfx.eu> wrote:
Hi Paul,
On 30 Jun 2013, at 19:38, Paul DeBruicker <pdebruic@gmail.com> wrote:
Hi -
I've got both Zinc and Zodiac loaded into a Pharo 2 image. When I do a ZnClient #get: to an http connection I'm getting the dnu mentioned in the subject of this email.
The message send that causes it is in ZnStringEntity>>#readFrom:
ZnNetworkingUtils default socketStreamClass prints: 'ZdcSocketStream'
Changing that to 'SocketStream' like so:
ZnNetworkingUtils default socketStreamClass: SocketStream
makes everything work .
Just wanted to let you know
Paul
Thanks a lot for reporting this issue.
It is a bit weird though: I normally don't use #position on principle, but apparently I did in some recent optimisation code, I also vaguely remember adding #position somewhere..
Furthermore, all unit tests are green on both 2.0 and latest 3.0 with ZdcSocketStream as default.
Step one would be to add a unit test.
Is it possible for you to give me the URL or a similar one exposing the problem - it would have to be related to non-ASCII, non-Latin-1 characters I guess.
Thanks,
Sven
-- Sven Van Caekenberghe Proudly supporting Pharo http://pharo.org http://association.pharo.org http://consortium.pharo.org
participants (3)
-
Paul DeBruicker -
Stéphane Ducasse -
Sven Van Caekenberghe