Thread: A starter task
Can you lead me to a beginner friendly task so I can start hacking?
Siavosh Kasravi
"Save a Tree" - Please print this email only if necessary.
Sorry I am not sure if I am doing this right. Should I look somewhere else for tasks?
On Fri, Sep 13, 2024 at 7:57 PM sia kc <siavosh.kasravi@gmail.com> wrote:
Can you lead me to a beginner friendly task so I can start hacking?--Siavosh Kasravi"Save a Tree" - Please print this email only if necessary.
Siavosh Kasravi
"Save a Tree" - Please print this email only if necessary.
On 9/15/24 21:42, sia kc wrote: > Sorry I am not sure if I am doing this right. Should I look somewhere > else for tasks? > Hi, I think you can take a look at https://wiki.postgresql.org/wiki/Todo and see if there's a patch/topic you would be interested in. It's really difficult to "assign" a task based on a single sentence, with no info about the person (experience with other projects, etc.). If you're staring with PostgreSQL development, maybe take a look at https://wiki.postgresql.org/wiki/Developer_FAQ first. There's a lot of ways to develop, but this is a good intro. FWIW, maybe it'd be better to start by looking at existing patches and do a bit of a review, learn how to apply/test those and learn from them. regards -- Tomas Vondra
I am reading the documents. Think the Todo list is what I needed thanks.
I think I should respond by sending my response to the mailing list but not sure why gmail does not have such a button. Please correct me if I am wrong.
On Sun, Sep 15, 2024 at 11:30 PM Tomas Vondra <tomas@vondra.me> wrote:
On 9/15/24 21:42, sia kc wrote:
> Sorry I am not sure if I am doing this right. Should I look somewhere
> else for tasks?
>
Hi,
I think you can take a look at https://wiki.postgresql.org/wiki/Todo and
see if there's a patch/topic you would be interested in. It's really
difficult to "assign" a task based on a single sentence, with no info
about the person (experience with other projects, etc.).
If you're staring with PostgreSQL development, maybe take a look at
https://wiki.postgresql.org/wiki/Developer_FAQ first. There's a lot of
ways to develop, but this is a good intro.
FWIW, maybe it'd be better to start by looking at existing patches and
do a bit of a review, learn how to apply/test those and learn from them.
regards
--
Tomas Vondra
Siavosh Kasravi
"Save a Tree" - Please print this email only if necessary.
Tomas Vondra <tomas@vondra.me> writes: > I think you can take a look at https://wiki.postgresql.org/wiki/Todo and > see if there's a patch/topic you would be interested in. It's really > difficult to "assign" a task based on a single sentence, with no info > about the person (experience with other projects, etc.). Beware that that TODO list is poorly maintained, so items may be out of date. Worse, most of what's there got there because it's hard, or there's not consensus about what the feature should look like, or both. So IMO it's not a great place for a beginner to start. > FWIW, maybe it'd be better to start by looking at existing patches and > do a bit of a review, learn how to apply/test those and learn from them. Yeah, this is a good way to get some exposure to our code base and development practices. regards, tom lane
On 9/15/24 22:10, sia kc wrote: > I am reading the documents. Think the Todo list is what I needed thanks. > I think I should respond by sending my response to the mailing list but > not sure why gmail does not have such a button. Please correct me if I > am wrong. > I'm pretty sure gmail has "reply to all" button somewhere. Also, please don't top post, reply in line. -- Tomas Vondra
On 9/15/24 22:43, Tom Lane wrote: > Tomas Vondra <tomas@vondra.me> writes: >> I think you can take a look at https://wiki.postgresql.org/wiki/Todo and >> see if there's a patch/topic you would be interested in. It's really >> difficult to "assign" a task based on a single sentence, with no info >> about the person (experience with other projects, etc.). > > Beware that that TODO list is poorly maintained, so items may be out > of date. Worse, most of what's there got there because it's hard, > or there's not consensus about what the feature should look like, > or both. So IMO it's not a great place for a beginner to start. > True, some of the items may be obsolete, some are likely much harder than expected (or perhaps even infeasible), etc. But it's the only such list we have, and it is at least a reasonable overview of the areas. Presumably a new contributor will start by discussing the patch first, and won't waste too much time on it. regards -- Tomas Vondra
So isn't there something like Jira backlog to manage the tasks?
My plan is to do some really small tasks like adding an option to a command but aim for harder ones like optimizer stuff.
On Mon, Sep 16, 2024, 00:14 Tom Lane <tgl@sss.pgh.pa.us> wrote:
Tomas Vondra <tomas@vondra.me> writes:
> I think you can take a look at https://wiki.postgresql.org/wiki/Todo and
> see if there's a patch/topic you would be interested in. It's really
> difficult to "assign" a task based on a single sentence, with no info
> about the person (experience with other projects, etc.).
Beware that that TODO list is poorly maintained, so items may be out
of date. Worse, most of what's there got there because it's hard,
or there's not consensus about what the feature should look like,
or both. So IMO it's not a great place for a beginner to start.
> FWIW, maybe it'd be better to start by looking at existing patches and
> do a bit of a review, learn how to apply/test those and learn from them.
Yeah, this is a good way to get some exposure to our code base and
development practices.
regards, tom lane
Tomas Vondra <tomas@vondra.me> writes: > Presumably a new contributor will start by discussing the patch first, > and won't waste too much time on it. Yeah, that is a really critical piece of advice for a newbie: no matter what size of patch you are thinking about, a big part of the job will be to sell it to the rest of the community. It helps a lot to solicit advice while you're still at the design stage, before you spend a lot of time writing code you might have to throw away. Stuff that is on the TODO list has a bit of an advantage here, because that indicates there's been at least some interest and previous discussion. But that doesn't go very far, particularly if there was not consensus about how to do the item. Job 1 is to build that consensus. regards, tom lane
I have a bad experience. I picked up a task from MariaDB backlog, explained in their chat rooms that I started doing that. After it was done which was a SQL command which MySQL already supported to restart server instance with SQL, they started rethinking the validity of the feature for the MariaDB. So the task got suspended.
About inlining not sure how it is done with gmail. Maybe should use another email client.
About reply to all button, I think only sending to mailing list address should suffice. Why including previous recipients too?
sia kc <siavosh.kasravi@gmail.com> writes: > About reply to all button, I think only sending to mailing list address > should suffice. Why including previous recipients too? It's a longstanding habit around here for a couple of reasons: * The mail list servers are occasionally slow. (Our infrastructure is way better than it once was, but that still happens sometimes.) If you directly cc: somebody, they can reply to that copy right away whether or not they get a copy from the list right away. * pgsql-hackers is a fire hose. cc'ing people who have shown interest in the thread is useful because they will get those copies separately from the list traffic, and so they can follow the thread without having to dig through all the traffic. regards, tom lane
On 9/16/24 00:32, Tom Lane wrote: > sia kc <siavosh.kasravi@gmail.com> writes: >> About reply to all button, I think only sending to mailing list address >> should suffice. Why including previous recipients too? > > It's a longstanding habit around here for a couple of reasons: > > * The mail list servers are occasionally slow. (Our infrastructure > is way better than it once was, but that still happens sometimes.) > If you directly cc: somebody, they can reply to that copy right away > whether or not they get a copy from the list right away. > > * pgsql-hackers is a fire hose. cc'ing people who have shown interest > in the thread is useful because they will get those copies separately > from the list traffic, and so they can follow the thread without > having to dig through all the traffic. > True, but it's also up to the client - the messages sent through the mailing list have the appropriate headers (List-Id etc.) and it's up to the client to show the "reply to list" button. Thunderbird does, for example. I thought gmail would too, but perhaps it doesn't. regards -- Tomas Vondra
FWIW, maybe it'd be better to start by looking at existing patches and
do a bit of a review, learn how to apply/test those and learn from them.
lets say i have experience in wal,physical replication,buffer management where can i find patches to review on these topics?
regards
Tony Wayne
do a bit of a review, learn how to apply/test those and learn from them.
lets say i have experience in wal,physical replication,buffer management where can i find patches to review on these topics?
regards
Tony Wayne
On Mon, Sep 16, 2024 at 1:30 AM Tomas Vondra <tomas@vondra.me> wrote:
On 9/15/24 21:42, sia kc wrote:
> Sorry I am not sure if I am doing this right. Should I look somewhere
> else for tasks?
>
Hi,
I think you can take a look at https://wiki.postgresql.org/wiki/Todo and
see if there's a patch/topic you would be interested in. It's really
difficult to "assign" a task based on a single sentence, with no info
about the person (experience with other projects, etc.).
If you're staring with PostgreSQL development, maybe take a look at
https://wiki.postgresql.org/wiki/Developer_FAQ first. There's a lot of
ways to develop, but this is a good intro.
FWIW, maybe it'd be better to start by looking at existing patches and
do a bit of a review, learn how to apply/test those and learn from them.
regards
--
Tomas Vondra
On 9/16/24 08:49, Tony Wayne wrote: > FWIW, maybe it'd be better to start by looking at existing patches and > do a bit of a review, learn how to apply/test those and learn from them. > > lets say i have experience in wal,physical replication,buffer management > where can i find patches to review on these topics? > Start by looking at the current commitfest: https://commitfest.postgresql.org/49/ This is the only place tracking patches people are currently working on, and submitted to the mailing list for a discussion/review. regards -- Tomas Vondra
On 9/16/24 00:17, sia kc wrote: > I have a bad experience. I picked up a task from MariaDB backlog, > explained in their chat rooms that I started doing that. After it was > done which was a SQL command which MySQL already supported to restart > server instance with SQL, they started rethinking the validity of the > feature for the MariaDB. So the task got suspended. > Unfortunately this can happen here too, to some extent. Sometimes it's not obvious how complex the patch will be, the feature may conflict with another feature in some unexpected way, etc. It's not like we have a 100% validated and agreed design somewhere. This is why my advice is to pick a patch the contributor is personally interested in. It puts him/her in a better position to advocate for the feature, decide what trade offs are more appropriate, etc. > About inlining not sure how it is done with gmail. Maybe should use > another email client. > Can you just expand the email, hit enter in a place where you want to add a response, and write. regards -- Tomas Vondra
On Mon, Sep 16, 2024 at 11:28 AM Tomas Vondra <tomas@vondra.me> wrote:
On 9/16/24 00:17, sia kc wrote:
> I have a bad experience. I picked up a task from MariaDB backlog,
> explained in their chat rooms that I started doing that. After it was
> done which was a SQL command which MySQL already supported to restart
> server instance with SQL, they started rethinking the validity of the
> feature for the MariaDB. So the task got suspended.
>
Unfortunately this can happen here too, to some extent. Sometimes it's
not obvious how complex the patch will be, the feature may conflict with
another feature in some unexpected way, etc. It's not like we have a
100% validated and agreed design somewhere.
This is why my advice is to pick a patch the contributor is personally
interested in. It puts him/her in a better position to advocate for the
feature, decide what trade offs are more appropriate, etc.
By picking a patch I assume you mean picking an already done task and seeing for example how I would have done it, right?
> About inlining not sure how it is done with gmail. Maybe should use
> another email client.
>
Can you just expand the email, hit enter in a place where you want to
add a response, and write.
Thanks.
Siavosh Kasravi
"Save a Tree" - Please print this email only if necessary.
On 9/16/24 10:35, sia kc wrote: > > > On Mon, Sep 16, 2024 at 11:28 AM Tomas Vondra <tomas@vondra.me > <mailto:tomas@vondra.me>> wrote: > > On 9/16/24 00:17, sia kc wrote: > > I have a bad experience. I picked up a task from MariaDB backlog, > > explained in their chat rooms that I started doing that. After it was > > done which was a SQL command which MySQL already supported to restart > > server instance with SQL, they started rethinking the validity of the > > feature for the MariaDB. So the task got suspended. > > > > Unfortunately this can happen here too, to some extent. Sometimes it's > not obvious how complex the patch will be, the feature may conflict with > another feature in some unexpected way, etc. It's not like we have a > 100% validated and agreed design somewhere. > > > > This is why my advice is to pick a patch the contributor is personally > interested in. It puts him/her in a better position to advocate for the > feature, decide what trade offs are more appropriate, etc. > > By picking a patch I assume you mean picking an already done task and > seeing for example how I would have done it, right? > I mean both the patch you'd review and the patch/feature you'd be writing yourself. My experience is that when a person is genuinely interested in a topic, that makes it easier to reason about approaches, trade offs, and stick with the patch even if it doesn't go smoothly. It's a bit similar to a homework. I always absolutely hated homework done only for the sake of a homework, and done the absolutely bare minimum. But if it was something useful/interesting, I'd spend hours perfecting it. Patches are similar, IMO. If you pick a patch that's useful for you (e.g. the feature would make your job easier), that's a huge advantage IMO. regards -- Tomas Vondra
On 2024-09-15 Su 6:17 PM, sia kc wrote:
About inlining not sure how it is done with gmail. Maybe should use another email client.
Click the three dots with the tooltip "Show trimmed content". Then you can scroll down and put your reply inline. (Personally I detest the Gmail web interface, and use a different MUA, but you can do this even with the Gmail web app.)
cheers
andrew
-- Andrew Dunstan EDB: https://www.enterprisedb.com