Copied as much as i could from Numerus, and made as few modifications as
i could to adapt to this code base; it is, quite frankly, a piece of
shit.
We need to be able to create invoices from scratch “just in case”,
apparently, but it is not yet possible to create an invoice from a
booking.
I need to retrieve the values from the database and put them in the
form, like all other forms, but in this case the processing is done as
if it were a new form, because everything comes from the query string
and there is no need to do any extra work then.
Had to move the <footer> from the fields.gohtml to form.gohtml because
then it could not know that it was editing an existing booking. Had to
move the <fieldset> out too, in order to give it an ID and make it
htmx’s target, or it would replace the form, causing even more problems
—the button would disappear then—. The target **must** be in <form>
because it is needed for tis children’s hx-get and for its own hx-put.
Had to bring the same fields that i have for a payment to booking,
except that some of those should be nullable, because it is unreasonable
to ask front desk to gather all customer data when they have a booking
via phone, for instance.
Therefore, i can not take advantage of the validation for customer data
that i use in the public-facing form, but, fortunately, most of the
validations where in separated functions, thus only had to rewrite that
one for this case.
I already have to create a booking from a payment, when receiving a
payment from the public instance, thus i made that function and reused
it here. Then i “overwrite” the newly created pre-booking with the
customer data from the form, and set is as confirmed, as we do not see
any point of allowing pre-bookings from employees.
This is actually only used for plots, but, of course, it means that
every booking now can potentially have many booked campsites, and have
to create a relation for it.
I now have a conundrum regarding stay dates: i need them to be in the
same table as the campsite_id, because constraints only work on a single
relation and without the dates i can not make sure that i am not
overbooking a given campsite; but, on the other hand, all campsites
under the same booking must be for the same dates.
Where does stay belong, then? In booking or booking_campsite? If in
booking then i can not have a constraint that most assuredly will bite
me in the back, but if in booking_campsite then each campsite could
potentially have different dates.
As far as i can see, i can not use a exclude constraint with <> for
dates in booking_campsite to ensure that all rows with the same
booking_id have the same stay (i.e., exclude those that have a different
stay for the same booking_id).
For now, the say is in **both** relations: in booking, because i need it
when it is a prebooking, at least, and in booking_campsite for the
aforementioned constraint requirements.
Will this come back and bite me? Yes, it will. But what can i do?
I need the campsite_id in booking to know what row to show the booking
at. Besides the need of knowing which actual campsite has been booked,
of course.
This field is nullable because we can not now it until an employee has
confirmed the booking; until that point we only know the campsite type
customer requested. I do not care much if the campsite_id is from a
different campsite_type, because maybe the customer requested the change
by phone or what have you, therefore the database can not be that
strict. It must have a value if the booking is confirmed.
It helps me if the arrival_date and departure_date is a single
daterange, because then i can use `&&` and other range operators to work
with these dates. For instance, i have to intersect it with the range
displayed on the screen in order to know which day i have to put it.
But then i have to know whether the booking begins and ends in the
display range, because i only have to show arrival and departure (i.e.,
the box half-way within the first or last boxes) on these days only.
It turns out that, **this time**, at least, the way to compute the
discount is not by “the more expensive”, but “the more expensive _in a
given group_”.
However, there are a couple of options, such as motorhome, that can be
in different groups but only must be used once.
I swear i believe sometime before we said that the number of dogs is not
important and should be used only as a boolean, but apparently it is
wrong: it should be number_dogs * cost_per_pet.
After months of keeping what does the ACSI checkbox mean, now customer
told us that we should add a discount based on a series of
arbitrary conditions that, and need to be done NOW!
There is no UI to edit the conditions due to lack of time.
I always intended to delete draft payments after some time. I follow
WooCommerce’s default times: 1 day for draft and 1 hour for pending. No
other reason than we are used to it.
I added a cron job, rather than a systemd timer, because i want email
notifications, and because i do not yet know how to add many service
files in a Debian package.
This is required by law.
I do not know why i have this value and the tax amount in the database,
but the payment percent and the number of days are hardcoded. I guess i
am such an inconsistent mess.
Had to do a couple of changes to the database: add the currency_code to
the payment relation, to format the price according to the payment’s
currency instead of the company’s; and the reference SQL function, to
replace the equivalent golang function, so that i can use it to index
payments.
The rest is mostly the same as any other page, except that the
individual payment’s page is not a form, but a regular info dump.
I also moved the payment settings as a sub-route of payments, as i
believe this makes more sense than an additional user menu item.
Customer wants to require a down payment of 30 % for bookings made
one week or more before the actual date, and to make the full payment
otherwise.
This would require yet another relation to keep these values. Fuck it;
i added them to the function, as they are very unlikely to change.
That forced me to change the test for draft_payment to use relative
dates, otherwise there is no way i can have stable results in the
future.
This is easier to read and requires less unit tests, but i only used
them in the new relations and fields for HEAD, because i do not see any
point on creating migrations just for that.
I have to basically do the reverse of signing the request to verify that
the notification comes from them. Lots of code just for that.
I return the changed status from the PL/pgSQL function because i will
need to email customers when a payment is completed, and i need to know
when.
I had to add the payment concept separate from the booking, unlike other
eCommerce solutions that subsume the two into a single “order”, like
WooCommerce, because bookings should be done in a separate Camper
instance that will sync to the public instance, but the payment is done
by the public instance. There will be a queue or something between
the public and the private instance to pass along the booking
information once the payment is complete, but the public instance still
needs to keep track of payments without creating bookings.
To compute the total for that payment i had to do the same as was doing
until now for the cart. To prevent duplications, or having functions
with complex return types, i now create a “draft” payment while the
user is filling in the form, and compute the cart there; from Go i only
have to retrieve the data from the relation, that simplifies the work,
actually.
Since the payment is computed way before customers enter their details,
i can not have that data in the same payment relation, unless i allow
NULL values. Allowing NULL values means that i can create a payment
without customer, thus i moved all customer details to a separate
relation. It still allows payment without customer, but at least there
are no NULL values.
Draft payments should be removed after a time, but i believe this needs
to be done in a cronjob or similar, not in the Go application.
To update the same payment while filling the same booking form, i now
have a hidden field with the payment slug. A competent developer would
have used a cookie or something like that; i am not competent.
Customer told us that there are some options, such as towels, that have
a fixed price for the whole stay, not a per night price. Thus, had to
add a boolean to know whether to use sum or max when computing the
cart’s total for each option.
It is a separate relation, instead of having a field in campsite_type,
because not all campsite types allow dogs. I could have added a new
field to campsite_type, but then its values it would be meaningless for
campsites that do not allow dogs, and a nullable field is not a valid
solution because NULL means “unknown”, but we **do** know the price —
none.
A separate relation encodes the same information without ambiguities nor
null values, and, in fact, removed the dogs_allowed field from
campsite_type to prevent erroneous status, such as a campsite type that
allows dogs without having a cost — even if the cost is zero, it has to
be added to the new relation.
I have to ask number and age ranges of hosts of guests for all campsite
types, not only those that have price options for adults, children, etc.
because i must compute the tourist tax for adults. These numbers will
be used to generate de rows for guests when actually creating the
booking, which is not done already.
To satisfy the campsite types that do have a price per guest, not only
per night, i had to add the prices for each range in the
campsite_type_cost relation. If a campsite type does not have price
per person, then that should be zero; the website then does not display
the price.
The minimal price for any campsite type is one adult for one night,
thus to compute the price i need at least the campsite type, the dates,
and the number of adults, that has a minimum of one. I changed the
order of the form to ask for these values first, so i can compute the
initial price as soon as possible. To help further, i show the
<fieldset>s progressively when visitors select options.
Customer told us that the minimum number of nights is per campsite type,
not per season. And he wants this, along with the maximum number of
nights, in order to limit the range of departure dates that guests can
choose when booking.
The “overflow” is for when people want to book plots for more guests
than is permitted, which the system would need to add a new plot to the
“shopping cart”, as it were; not implemented yet.
The ask zone preferences is to whether show the corresponding input on
the booking form, that it was done implicitly when the campsite type had
options, because up until now it was only for plots, but it is no longer
the case, thus i need to know when to show it; now it is explicit.
This is more or less the same as the campsites, as public information
goes, but for buildings and other amenities that the camping provides
that are not campsites.
A small page with a brief description, carousel, and feature list of
each individual accommodation.
Most of the relations and functions for carousel and features are like
the ones for campsite types, but i had to use the accommodation’s label
to find them, because they do not have slugs; i did not even though
these would be public, and they already have a label, although not
unique for all companies, like UUID slugs are.
We were using what was very clearly a campfire as the icon for the
barbecue service, when we first replaced it with a Font Awesome icon
we chose an icon that was neither campfire nor barbecue (flame-curved),
but now Oriol wanted one that left no doubt it was a barbecue.
Instead of replacing the campfire SVG with that of the barbecue, i have
chosen a campfire image from Font Awesome for our icon, and added a
separate icon for that service.
Apparently, each campsite type could have different check-in and
check-out times, thus i need them in the database.
I thought about using an integer or a datetime field, but customer seems
to want a text field to maybe add “before” and “after” there as well.
Translatable text it is.
There is no way, for now, to add, edit or remove users, because
currently we only need to list users.
I can not give admins access to the user table, for security
permissions, so i had to create a new view. I could name it also ‘user’
in ‘camper’ scheme, but then i was afraid i would have problems with
unit tests and their search_path, so instead i called it
‘company_user_profile’, which is like ‘user_profile’ but for all users
in ‘company_user’.
I created a new Go package for it, rather than add the admin handler in
‘auth’, because ‘template’ depends on ‘auth’, and rendering from ‘auth’
would cause a dependency loop.
I needed to have the roles in gettext to translate them, but there is
no obvious place where to put the call to PgettextNoop. For now, there
are in ‘NewAdminHandler’ because it is called once in the application’s
lifetime and they actually do not matter much.
This is a separate carousel from the one displayed at the bottom with
location info; it is, i suppose, a carousel for the hero image.
For the database, it works exactly as the home carousel, but on the
front had to use AlpineJS instead of Slick because it needs to show a
text popping up from the bottom when the slide is show, something i do
not know how to do in Slick.
It now makes no sense to have the carousel inside the “nature” section,
because the heading is no longer in there, and moved it out into a new
“hero” div.
Since i now have two carousels in home, i had to add additional
attributes to carousel.AdminHandler to know which URL to point to when
POSTing, PUTting, or redirecting.
Customer does not want the new “masonry-like” design of the surroundings
page, and wants the same style they already had: a regular list with
text and photo, alternating the photo’s side.
And, of course, they want to be able to add and edit them themselves. It
is like another carousel, but with an additional rich-text description.
The photos that we had in that page are no longer of use.
Customer does not want a contact page, but a page where they can write
the direction on how to reach the campground, with a Google map embed
instead of using Leaflet, because Google Maps shows the reviews right
in the map.
That means i had to replace the GPS locations with XML fields for the
customer to write. In all four languages.
This time i tried a translation approach inspired by PrestaShop: instead
of opening a new page for each language, i have all languages in the
same page and use AlpineJS to show just a single language. It is far
easier to write the translations, even though you do not have the source
text visible, specially in this section that there is no place for me
to put the language links.
I use Sortable, exactly like HTMx’s sorting example does[0]. Had to
export the slug or ID of some entries to be able to add it in the hidden
input.
For forms that use ID instead of slug, had to use an input name other
than “id” because otherwise the swap would fail due to bug #1496[1]. It
is apparently fixed in a recent version of HTMx, but i did not want to
update for fear of behaviour changes.
[0]: https://htmx.org/examples/sortable/
[1]: https://github.com/bigskysoftware/htmx/issues/1496
Every company need to have its own merchant code and encryption key,
thus it is not possible to use environment variables to keep that data,
and i have to store it in the database.
I do not want to give SELECT permission on the encryption key to guest,
because i am going to fuck it up sooner or later, and everyone would be
able to read that secret; i know it would. Therefore, i need a security
definer function that takes the data to encrypt, use the key to encrypt
it, and returns the result; nobody else should have access to that key,
not even admins!
By the way, i found out that every merchant receives the same key, thus
it is not a problem to keep it in the repository.
Since i need that SQL function to encrypt the data, i thought that i may
go the whole nine yards and sign the request in PostgreSQL too, after
all the data to sign comes from there, and it has JSON functions to
create and base64-code an object.
Fortunately, pg_crypto has all the functions that i need, but i can no
longer keep that extension inside the auth schema, because it is used
from others, and the public schema, like every other extensions, seems
more appropriate.
Instead of having the list of currency and language codes that Redsys
uses as constants in the code, i moved that as field to the currency
and language relations, so i can simply pass the lang_tag to the
function and it can transform that tag to the correct code; the currency
is from the company’s relation, since it is the only currency used in
the whole application (for now).
As a consequence, i had to grant execute to currency and the parse_price
functions to guest, too.
To generate the test data used in the unit tests, i used a third-party
PHP implementation[0], but i only got from that the resulting base64-coded
JSON object and signature, using the same that as in the unit test, and
did not use any code from there.
PostgreSQL formats the JSON as text differently than most
implementations i have seen: it adds spaces between the key name and
the colons, and space between the value and the separating comma. The
first implementation used replace() to format the JSON as exactly as
the PHP implementation, so that the result matches, and then tried to do
generate the form by hand using the output from PostgreSQL without the
replace(), to verify that Redsys would still accept my input. Finally,
i adjusted the unit test to whatever pg_prove said it was getting from
the function.
I still have the form’s action hard-codded to the test environment, but
the idea is that administrators should be able to switch from test to
live themselves. That means that i need that info in the redsys
relation as well. I think this is one of the few use cases for SQL’s
types, because it is unlikely to change anytime soon, and i do not need
the actual labels.
Unfortunately, i could not use enumerations for the request’s
transaction type because i can not attach an arbitrary number to the
enum’s values. Having a relation is overkill, because i would need
a constant in Go to refer to its primary key anyway, thus i kept the
same constant i had before for that. Language and currency constant
went out, as this is in the corresponding relations.
In setup_redsys i must have a separate update if the encrypt_key is null
because PostgreSQL checks constraints before key conflict, and i do
not want to give a default value to the key if the row is not there yet.
The problem is that i want to use null to mean “keep the same password”,
because it is what i intend to do with the user-facing form: leave the
input empty to keep the same password.
As now Go needs to pass composite types back and forth with PostgreSQL,
i need to register these types, and i have to do it every time because
the only moment i have access to the non-pooled connection is in the
AfterConnect function, but at that point i have no idea whether the
user is going to request a payment. I do not know how much the
performance degrades because of this.
[0]: https://github.com/ssheduardo/sermepa/blob/master/src/Sermepa/Tpv/Tpv.php
jordi fita mas