diegogomezdeck@consultar.com wrote:
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) Content-type: text/plain; charset=iso-8859-1 Content-transfer-encoding: 7BIT
New trial.
A few comments I have -
- I think it's usually better to separate tests out more into separate
methods, using setUp to create the required common fixtures (commonly used values, held in instance variables). That way it's more obvious what is broken, and it's better documentation of the intention of the test.
It's a subjective point. The test I submitted has a "line of thinking" from 2 equals objects to 2 different ones. In this case I prefer to show all the posibilities in 1 method.
Anyway I splited the test in 2 tests: #testEquals and #testNotEquals.
- #hash is strangely formatted.
I use prettyprint. But the prettyprint option don't store the code after changing it but before. So you see the formatting but it's not stored.
Anyway I create another version of #hash.
- It's probably better if people separate the Test changeset from the
fix changeset. This makes a few things easier -
- we can load the test before the fix, to see it break, and then load
the fix.
- they can easily be harvested separately (test now - fix after rework,
or test into test package - fix into update).
Done!
Note that the fix works just fine, came with a test, and is simple - it's generally very good, needs just a little polish. Most of the reason it's rejected is that it's important to raise our awareness of what makes work for the harvesters, so that future stuff is nearer perfect.
Let's try now.
Thanks, Diego. Daniel
Diego Gomez Deck
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) Content-type: application/x-gzip-compressed; name=MethodReferenceHashTest-dgd.2.cs.gz Content-transfer-encoding: base64 Content-disposition: attachment; filename=MethodReferenceHashTest-dgd.2.cs.gz
H4sIAAAAAAAAAM1VTW8aMRA9++D/MNDDJlIh4SMqQg0qhaa9JIcQ9RKhyuzOsqusbWJ7S1Plx3ds aEJgqRqlanuzZ8dvnt971kZnRkuY3JYobjrNLugUWnAuTJxB+/i4A9eFcGgdlIuEFn14ddJ6czwF raC32SYctFr9k26/3QYhoxqvjzKh5ggTdH3GztFlOrnEFA2qGD8Jm10RaiOZJ3zscRljm3h8WNIB 48vjHOcaPmqJ32GM8Q28DZVQ8Pt3I61sWbihacZaDjhfzYKtkX7cYHAKuYVcLgqUqBwmMCtddWdG FKlZRY7zRnWL1+XDbSkKO1SJvxJAkMplpIbN9NLSEgGN0QYOhEpgYfQMQzHNvx3yeo1zDzUSFsGW s7gQ1pLEFdM4y5V1gnafhcnFrMALIZGao4izcK6ivtC6GOexy7Wijw/dpPdcmzvaXVGDbbw3RBUN mcZrFaNBhpo9Iz8guiyVytU8AmIjF1QgC6Fz1DsKMaAMdHsE9CgNZ/cgtlBBKE0imO3yPWc7rV+2 tQeFyyY1VkPsa68DZ26pITXofV1rab1NZUExIJsCXVq5JaLyJsk6ZxaLlDO2auvD9Q6/0z2XmTZ/ 43AI2T6E8HW6wf2RtcEF3YMSTE6EOFkyfe0THIRNyAT40NHpUCE2GDttDncuXa/Q3aKb0LREmGS0 iuXEGT9uNWVyJ2e6oKymWu9343kg/1xrqMEffgMX+j97Bn/f6lUQK6F8sJM8DRBuHVm0r+FJXKEx AKVBarMR1yduK/fcsNR/vomHkURjzeKXc1+Y85kwzZeRX0W0KqEGpf6KkzV7Grbze6rxH/JXhDDt BwAA
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) Content-type: application/x-gzip-compressed; name=MethodReferenceHashFix-dgd.2.cs.gz Content-transfer-encoding: base64 Content-disposition: attachment; filename=MethodReferenceHashFix-dgd.2.cs.gz
H4sIAAAAAAAAAKVTTW/aQBA9+7D/YUwOTg4hOAQFWQUVBdFySCuVHipFVFrsAW/j9bq7ayj99Z1d 0wYoKJV6sXbH783He7PRRCsJs+818udu+w7UEmJ45DrN4bbT6cJTwS0aC3WV0SGBi15835mDKqG/ D+MW4jjp3SW3PeAyClnrIeflCmGGNgmCR7S5yj7hEjWWKb7nJp+IH9fZKmNjlzYIgv10bFQTXrvw WOBKwTsl8SeMMX2GNz7iA+7+9kGVpi7sSLdTJYeMNaXgqOJnmmE4HIAwIGRVoMTSYgaL2p5G5tQh gcvIMnb9QuFW0OSk0VkSoTX+I34AG2FzGH0YU1elscgzh51OjM8i1RpPE02FqUADl8JGBkplocQU DddbWGCqaoMuz8fFN0xf0FesFTLGwqOMIP3dTEhviEjDimthSNQIjOWyoiDZBN2b+xvvdNxN4j4Z PABOhXPUx/lY0BqVZoMaNjk6ANCHxklRrOnCy8wHuF7VTiL6U2k07uTCLDBcUvu+93aLBV/BYLGE 3yMPztbdIVgQUIkEnjwtLbgxs61cqALOMvdA8xP0qSECf4XegI7ojbSvld9HzUMI/8Oj/s4j9xTJ I7+Qf+zgMJO8KKa0+SvvjqI9WfOiRvcsNLqXTt6oA79ov8gTkZE/wm4bQy4PHPFPZSHsF9fcX6K7 v1ekygHgQBeH8FMfD928gBkWtAmOerGrGLJfl5irZrUEAAA=
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) MIME-version: 1.0 Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7BIT
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw)--
I also added the test as a separate update. I'm not sure if I really need to do this, since the tests are being maintained in Marcus' package, but I incorporated it for now.
- Doug
On Saturday, March 8, 2003, at 06:51 AM, Daniel Vainsencher wrote:
diegogomezdeck@consultar.com wrote:
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) Content-type: text/plain; charset=iso-8859-1 Content-transfer-encoding: 7BIT
New trial.
A few comments I have -
- I think it's usually better to separate tests out more into
separate methods, using setUp to create the required common fixtures (commonly used values, held in instance variables). That way it's more obvious what is broken, and it's better documentation of the intention of the test.
It's a subjective point. The test I submitted has a "line of thinking" from 2 equals objects to 2 different ones. In this case I prefer to show all the posibilities in 1 method.
Anyway I splited the test in 2 tests: #testEquals and #testNotEquals.
- #hash is strangely formatted.
I use prettyprint. But the prettyprint option don't store the code after changing it but before. So you see the formatting but it's not stored.
Anyway I create another version of #hash.
- It's probably better if people separate the Test changeset from
the fix changeset. This makes a few things easier -
- we can load the test before the fix, to see it break, and then load
the fix.
- they can easily be harvested separately (test now - fix after
rework, or test into test package - fix into update).
Done!
Note that the fix works just fine, came with a test, and is simple - it's generally very good, needs just a little polish. Most of the reason it's rejected is that it's important to raise our awareness of what makes work for the harvesters, so that future stuff is nearer perfect.
Let's try now.
Thanks, Diego. Daniel
Diego Gomez Deck
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) Content-type: application/x-gzip-compressed; name=MethodReferenceHashTest-dgd.2.cs.gz Content-transfer-encoding: base64 Content-disposition: attachment; filename=MethodReferenceHashTest-dgd.2.cs.gz
H4sIAAAAAAAAAM1VTW8aMRA9++D/ MNDDJlIh4SMqQg0qhaa9JIcQ9RKhyuzOsqusbWJ7S1Plx3ds aEJgqRqlanuzZ8dvnt971kZnRkuY3JYobjrNLugUWnAuTJxB+/ i4A9eFcGgdlIuEFn14ddJ6czwF raC32SYctFr9k26/ 3QYhoxqvjzKh5ggTdH3GztFlOrnEFA2qGD8Jm10RaiOZJ3zscRljm3h8WNIB 48vjHOcaPmqJ32GM8Q28DZVQ8Pt3I61sWbihacZaDjhfzYKtkX7cYHAKuYVcLgqUqBwmMC tddWdG FKlZRY7zRnWL1+XDbSkKO1SJvxJAkMplpIbN9NLSEgGN0QYOhEpgYfQMQzHNvx3yeo1zDz USFsGW s7gQ1pLEFdM4y5V1gnafhcnFrMALIZGao4izcK6ivtC6GOexy7Wijw/ dpPdcmzvaXVGDbbw3RBUN mcZrFaNBhpo9Iz8guiyVytU8AmIjF1QgC6Fz1DsKMaAMdHsE9CgNZ/ cgtlBBKE0imO3yPWc7rV+2 tQeFyyY1VkPsa68DZ26pITXofV1rab1NZUExIJsCXVq5JaLyJsk6ZxaLlDO2auvD9Q6/ 0z2XmTZ/ 43AI2T6E8HW6wf2RtcEF3YMSTE6EOFkyfe0THIRNyAT40NHpUCE2GDttDncuXa/ Q3aKb0LREmGS0 iuXEGT9uNWVyJ2e6oKymWu9343kg/1xrqMEffgMX+j97Bn/ f6lUQK6F8sJM8DRBuHVm0r+FJXKEx AKVBarMR1yduK/fcsNR/ vomHkURjzeKXc1+Y85kwzZeRX0W0KqEGpf6KkzV7Grbze6rxH/JXhDDt BwAA
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) Content-type: application/x-gzip-compressed; name=MethodReferenceHashFix-dgd.2.cs.gz Content-transfer-encoding: base64 Content-disposition: attachment; filename=MethodReferenceHashFix-dgd.2.cs.gz
H4sIAAAAAAAAAKVTTW/aQBA9+7D/YUwOTg4hOAQFWQUVBdFySCuVHipFVFrsAW/ j9bq7ayj99Z1d 0wYoKJV6sXbH783He7PRRCsJs+818udu+w7UEmJ45DrN4bbT6cJTwS0aC3WV0SGBi15835 mDKqG/ D+MW4jjp3SW3PeAyClnrIeflCmGGNgmCR7S5yj7hEjWWKb7nJp+IH9fZKmNjlzYIgv10bF QTXrvw WOBKwTsl8SeMMX2GNz7iA+7+9kGVpi7sSLdTJYeMNaXgqOJnmmE4HIAwIGRVoMTSYgaL2p 5G5tQh gcvIMnb9QuFW0OSk0VkSoTX+I34AG2FzGH0YU1elscgzh51OjM8i1RpPE02FqUADl8JGBk plocQU DddbWGCqaoMuz8fFN0xf0FesFTLGwqOMIP3dTEhviEjDimthSNQIjOWyoiDZBN2b+xvvdN xN4j4Z PABOhXPUx/ lY0BqVZoMaNjk6ANCHxklRrOnCy8wHuF7VTiL6U2k07uTCLDBcUvu+93aLBV/BYLGE 3yMPztbdIVgQUIkEnjwtLbgxs61cqALOMvdA8xP0qSECf4XegI7ojbSvld9HzUMI/8Oj/ s4j9xTJ I7+Qf+zgMJO8KKa0+SvvjqI9WfOiRvcsNLqXTt6oA79ov8gTkZE/ wm4bQy4PHPFPZSHsF9fcX6K7 v1ekygHgQBeH8FMfD928gBkWtAmOerGrGLJfl5irZrUEAAA=
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw) MIME-version: 1.0 Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7BIT
--Boundary_(ID_lQW/mjyUZyicq8ba1d9JRw)--
On Wed, Apr 02, 2003 at 12:42:47AM -0500, Doug Way wrote:
I also added the test as a separate update. I'm not sure if I really need to do this, since the tests are being maintained in Marcus' package, but I incorporated it for now.
Problem with this is that I need to update the test-removal package each time... As soon as both tests-removal and sunit-removal are in the image, adding tests outside the testing package won't be possible.
Marcus
Marcus Denker wrote:
On Wed, Apr 02, 2003 at 12:42:47AM -0500, Doug Way wrote:
I also added the test as a separate update. I'm not sure if I really need to do this, since the tests are being maintained in Marcus' package, but I incorporated it for now.
Problem with this is that I need to update the test-removal package each time...
Okay, I won't add further tests directly to the image then.
Hmm, I guess I will have to somehow notify you when a fix with a test is being incorporated, so that you can add the test to your package. Or you will have to keep track of these [update] messages and check for fixes which have corresponding tests.
As soon as both tests-removal and sunit-removal are in the image, adding tests outside the testing package won't be possible.
True. :-)
- Doug Way
On Wed, Apr 02, 2003 at 07:01:16PM -0500, Doug Way wrote:
Hmm, I guess I will have to somehow notify you when a fix with a test is being incorporated, so that you can add the test to your package. Or you will have to keep track of these [update] messages and check for fixes which have corresponding tests.
I add all tests that are sent with the [TEST] tag, even without having a fix.
The idea is that the failing testcases can be some kind of "guide" or todo-list as soon as we have a testing server (someone's working on that). SUnit 3.1 added logging, so the failing test can reference the fix or bug in the sqfixes archive.
Marcus
squeak-dev@lists.squeakfoundation.org